Drivers of Fish Diversity and Turnover Across Multiple Spatial Scales: Implications for Conservation in the Western Ghats, India

Home , Bhavania, Chalakudy River, Dawkinsia, Dawkinsia filamentosa, Devachihatti, Gangamoola

Drivers of fish diversity and turnover across multiple spatial scales: Implications for conservation in the Western Ghats, India

Thesis Submitted for the Degree of Doctor of Philosophy

Thesis Submitted By Vidyadhar Atkore

Reg. No: 110900028

Thesis Submitted to Manipal University Manipal, 576104

Research carried out at The Academy for Conservation Science and Sustainability Studies Ashoka Trust for Research in Ecology and the Environment (ATREE) Bangalore

Declaration by the candidate

I, Vidyadhar Atkore hereby declare that the thesis entitled “Drivers of fish diversity and turnover across multiple spatial scales: Implications for conservation in the Western Ghats, India” has been compiled by me under the supervision of Dr. Jagdish Krishnaswamy, Ashoka Trust for Research in Ecology and the Environment (ATREE), Bangalore. The thesis has not been previously submitted for the award of any degree, diploma, associateship, fellowship, or its equivalent to any other University or Institution.

(Vidyadhar Atkore)

Place: Bangalore

Date: 30th November, 2017

Certificate

This is to certify that the thesis entitled “Drivers of fish diversity and turnover across multiple spatial scales: Implications for conservation in the Western Ghats, India” submitted by Mr. Vidyadhar Atkore, for the award of Doctor of Philosophy for, Manipal University, Manipal, is a record of the research work carried out by him during the period of his study in this university under my guidance and supervision and the thesis has not formed the basis for the award of any degree, diploma or other similar titles.

Dr. Jagdish Krishnaswamy Place: Bangalore Date: 30th November, 2017

Acknowledgements

I would like to express my sincere gratitude to several institutions and individuals who have helped me in finishing my doctoral research work.

Bangalore has been a perfect place to pursue my PhD. I was really blessed to learn and work with some of the best scientist and teachers in the world especially Drs. Jagdish Krishnaswamy, Sharad Lele, Nitin Rai, Suhel Quader, Kavita Isvaran, Kartik Shanker, T. Kannan, Harini Nagendra, Bijoy Thomas, Shrinivas Badiger, Seema Puroshottam, R. Ganesan, Soubadra Devi, Siddhartha Krishnan, Siddappa Setty, Ankila Hiremath, T. Ganesh, Veena Srinivasan, Ulka Kelkar, Abi Tamim Vanak, Maria Thaker and Anindya Sinha. Special thanks to several friends for their constant support and encouragement. The list is very long but I thank Obaiah, Prashant, Rakesh, Jayalakshmi, Vindhya, Drs. Arundhati Das, Susan Varghese, Paramesha Mallegowda, Jagadish M.R., Milind Buyan, Vivek Ramachandran, Chetan H.C. Office and administrative staff of ATREE have been very extremely supportive. My sincere thanks goes to Gopi, (late) Sindhu, Vartika, Ramesh, Saji, Laxminarayan, Raghu, Madhavi, Bhogaiah, Usha, Smita, Goutham and Indrani.

First and foremost I express my sincere gratitude to Dr. Jagdish Krishnaswamy, for readily agreeing and accepting me as his PhD student. He has been my role model and ideal scientist to work with. His illustrative academic career has always and continues to inspire me. He has immeasurably helped me throughout all the crucial stages of my research. I sincerely thank him for his able guidance and patience and giving me full freedom to work on my ideas and in my capacity. I had very rewarding and rich learning experience with him.

Dr. Kartik Shanker, my Doctoral Advisory Committee (DAC) member and informal co-supervisor has fostered my interest in community ecology and biogeography. He has been instrumental in shaping my research question and provided necessary lab space in CES, IISc. Dr. Shrinivas Badiger another DAC member has helped me in finalizing and securing funding for one of the field site. I was benefited from his inputs on the study design and several useful comments on my synopsis and research articles. He and Dr. Jagdish guided me to present part of my research work at 4th Bio

Asia Seminar held at Chinese Academy of Science, Beijing – my first ever visit to any foreign country! Dr. Ajithkumar from Centre for Wildlife Studies, Bangalore (DAC member) has always been another source of inspiration. He has been very motivating and supportive throughout and provided several useful comments on my work. I also would like to express my sincere gratitude to former ATREE Directors, Dr. Gladwin Joseph and Dr. Ganesan Balachander for both academic and financial support at crucial stages of my PhD.

I am also indebted to Drs. Nitin Rai, Aravind Madhystha, Siddhartha Krishnan and Milind Buyan (Academic coordinators) and Ms. Madhvi Latha for all the academic support they have given during my PhD. I would like to thank Drs. G. Ravikant and Priyanka Jamwal for providing lab space for conducting water quality analysis. I also thank Manipal University for all the necessary Academic support. Both the examiners of my thesis are gratefully acknowledged. I was immensely benefited from their several positive and constructive comments.

Dr. M. Arunachalam was very kind enough to share literature on fish ecology from India. Dr. J.A. Johnson from Wildlife Institute of India, Dehradun gave me few important tips on field work - for instance, sketching a sampled stretch and noting down behavior of fish species. Dr. Anuradha Bhat has guided me on choosing fish sampling techniques and involved in very useful discussion on several occasions. Drs. Ranjit Daniels, C. P. Shaji, Markus Knight, Rema Devi madam and Rajeev Raghavan have helped me in identifying fishes. I was fortunate to meet Dr. K.C. Jayaram at his house in Chennai who not only encouraged me but also advised me to explore difficult terrain.

I would not have pursued my research without unconditional support and love from my friends. Foremost I would like to thank Dr. Rasheed Raza, Dr. Raman Kumar for encouraging me to pursue my dream. My lifelong friends, Hari, Rohit, Madhura, Anirban, Chandrima, Tamma, Amit, Abhishek, Mousumi, Joy, Karpa, Rishi, Barkha, Urbashi, Aniruddha, Radhika, Tamo, Navendu, Sneha, Divya, Nibedita, Nilambari, Ashok, Vijay, Varun, and Sandeep have been real source of inspiration due to their outstanding research.

Hari, Navendu, Sandeep, Tamma & Madhura have strengthened my chapters and manuscripts at different stages. I am very grateful to Sandeep Pulla, Krishnapriya Tamma, and Geetha Ramaswami for solving my statistical doubts. Several friends from abroad who helped me procuring equipment’s, and books especially I thank Dr. Sandeep Sharma, Dr. Harshad Kulkarni, Allwin Jesudasan, Raju dada & Smita Deshpande, David Manski and (late) Archana Bali.

I would also like to express my gratitude to Nachiket Kelkar for inviting me to co- conduct a workshop on Freshwater Biodiversity at the Student Conference on Conservation Science at Bangalore for 2013-2014 as well as requesting me to teach M. Sc. (Wildlife Biology and Conservation) of CWS-NCBS, Bangalore. I feel equally honored to seek his intellectual help in academics especially in the analysis and writing stage and being my collaborator. He has been always ready to teach and correct me on numerous occasions. I remain indebted to him for his kind help. Equally I am thankful to Dr. Mauricio Cetra from Brazil, Dr. Jani Heino from Finland, Dr. Maria Dornelas from UK, Dr. Sebastein Brosse and Dr. Pablo Tedesco from France for providing statistical advice and showing keen interest in my research.

Dr. Madhura Nipahdkar has taught me all the necessary techniques of GIS and offered her help whenever I needed. She has always believed in my abilities and constantly encouraged me to reach top level despite all hurdles. Sheethal V. R. has helped me in data organization. Susan for all the cheerful time and laughter in the lab. Zuhail, Vindhya and Praveen helped in the water analysis. Several volunteers on various occasion visited me in the field and helped me in data collection. These includes Ramprasad, Neethi, Ramya, Nisarg, Suman and Anoop. Several other friends including Ovee, Venkat, Vikram, Manish, Shivani, Kadambari, Rahul, Sandeep, Neeta, Shivona, and Tarun have taken part in discussion on several occasions and helped me at different stages. Ajay, Sarvanan and Yogisha for accommodating me in the field stations (Heggarne in Sirsi and Emerald in Nilgiri) during my thesis writing. I would also like to express my thanks to Gubbi Labs, Bangalore for requesting me to assess fish diversity in Sakleshpur taluk of Hassan district. I enjoyed working with Drs. H.S. Sudhira, K.V. Gururaja and their colleagues.

I thank the Forest Departments of Karnataka and Goa states for granting me research

iii permission to conduct my field study in one of the most beautiful landscape/s of the Western Ghats. I would like to thank Mr. Netalkar and Dr. Ramesh Kumar (DCF’s of Kudremukh Forest Division), Mr. T. L. Prahlad and Srikanth Kadolkar (RFO, of Kudremukh and Kerekatte wildlife range), and all the administrative staff especially, Mr. Jaynarayanan, Mr. Hariprasad of Kudremukh Divisional Forest, Karkala for their valuable support during the fieldwork. The frontline staff at Kudremukh was always ready to help me in providing the logistical support, I sincerely express my gratitude to Mr. Chandre gowda, Mr. Laxman ji, Seshappa for being my field companions and teaching me Kannada throughout my fieldwork without them I would not have explored and sample streams from the very remote and interior parts of Kudremukh NP. I would also like to thank Nagraj ji, Ramesh ji, Satish, Raghavendra, Manjunath, Umesh, Pramod, and Sandeep for helping me during different stages of my fieldwork. Kudremukh Iron Ore Limited (KICOL) at Kudremukh made my stay comfortable in Sahyadri guest house initially and lending me a beautiful house in Company quarters in Kudremukh town. All the administrative staff of the KICOL Company was very down to earth. I sincerely thank Manager, Mr. Ravindra (Surveyor, KIOCL Co. Ltd) and many resident doctors who showed a great interest in my research. I enjoyed talking and sharing my field experiences with Principal, teacher and students of the Kendriya Vidyalaya, Kudremukh. Amma who run a small tea stall called “The Hotel Kudremukh” who always served me fresh & hot food. The local bus transport service in Kudremukh is one of the best bus service that I ever enjoyed so far they are best at their timings.

Forest department of Goa and associated staff was very supportive. I sincerely thank DFO, Mr. Debendra Dalai Kamuprakash sir (ACF), Pranay Naik and Paresh Parab (RFO of Mollem and Mhadei Wildlife Sanctuaries & NP) and many friends from Goa - Sarvesh Panchwadkar, Devidas Kotkar, Rohit Gaokar, Yeshwant Gawas, Madhura, Amit Bandekar, Pankaj, Nirmal, Sachin and Parag Ranganekar for their continued support during my fieldwork.

I would like to thank Dr. T.V. Ramchandra, CES, IISc for allowing me to use his lab space and learn basics of water quality techniques where I became a friends with Alaknanda, Supriya, Durga and Sudarshan who not only taught me how to conduct water analysis but also were ready to answer my queries. Deepa from KS lab at CES,

IISc also helped me completing administrative formalities. Similarly, Dr. B. Venkatesh and Dr. B. Purandara from National Institute of Hydrology regional Centre at Belgaum facilitating water analysis in their lab. My comfortable stay in Khanapur was made possible by Ishwar Gowdapatil and Mrs. Khanapurkar that allowed me to complete my fieldwork on different times of the year.

Finally, I would like to express my sincere gratitude to my funders especially, Ecosystem Services for Poverty Alleviation (ESPA), Inlaks Shivdasani, Mumbai’s Ravi Sankaran Fellowship (2012), International Foundation for Science, Sweden, a grant from the Changing Water Cycle Program of Ministry of Earth Sciences, Government of India, Royal Norwegian Embassy (RNE) and National Mission for Himalayan Studies (NMHS).

This work would not have been completed without my loving parent’s tireless support and belief in anything and everything that I do. My siblings, especially, Yashodhara, Varsha, Sadanand and Dayanand for silently embracing all the hardship all these years and waiting for me to finish my PhD as soon as possible, I would be indebted to them rest of my life for their perseverance and patience.

Dedication I would like to dedicate this work to last remaining free- flowing tropical streams and rivers and their associated incredibly rich aquatic life

Executive Summary

Freshwater ecosystems are extraordinarily rich in their aquatic life. They sustain other terrestrial life and offer numerous ecosystem services to humankind. Fish is one of the most important components of freshwater ecosystems (Darwall et al., 2008). However, the status of freshwater habitat and associated aquatic life including that of fishes is in danger due to intense anthropogenic impacts such as small and big dams, water diversion schemes for irrigated agriculture, deforestation, removal of riparian cover, sand mining and pollution. As a result, freshwater fishes are being driven to extinction. The extinction risk for riverine fishes is believed to be far greater than terrestrial organisms, and may even have exceeded the natural rate of overall species extinction (Ricciardi and Rasmussen, 1999; Dias et al., 2017).

Studies that quantify various anthropogenic threats to biodiversity of the Tropical Asian streams and rivers are inadequate as compared to those in the temperate regions. Streams are being fragmented, disconnected and exploited heavily for rising human developmental needs. Ecologists have emphasized the need for their restoration, quantification of such threats and detail mapping of aquatic biodiversity (Strayer et al., 2010; Dudgeon et al., 2010; 2011; Araujo et al., 2013; Alexandre et al., 2013; Shimadzu et al., 2013; Sakaris, 2013; Bae et al., 2016). Therefore there is a need to address fish responses to the disturbances to the habitat at different spatial scales. Often lack of crucial information on species distribution, habitat ecology and species responses to different anthropogenic threats at multiple spatial scales impedes our ability to prioritize river conservation.

My PhD research integrates some of these ideas and quantifies the native fish diversity and factors that drive this diversity in four river sub-basins of the Western Ghat biodiversity hotspot in India. The study primarily assesses the fish diversity at multiple nested spatial scales i.e. segment, habitat, stream order, and sub-basin. Two river basins (Mhadei and Malaprabha) have numerous hydrological barriers in the form of small check dams, impoundments, barrages as well as other forms of disturbances such as substrate mining, fishing, and pollution from domestic and agricultural uses of rivers. Such disturbances have severely degraded the water quality and fish health in India (Daniels, 2002). The other two river sub-basins (Bhadra and

vii

Tunga) are hydrologically less modified but have local disturbances such as water diversion for agriculture, plantations and pollution.

Therefore the specific research questions were: 1). How do fish species diversity (richness and abundance) vary in different river segments in four sub-basins? 2). What are the drivers of fish species turnover between river segments in a sub-basin and across adjacent river sub-basin? 3). How do fish guild richness and composition vary across regulated and non-regulated sub-basins? and 4). What is the potential for fish species recovery downstream of hydrological barriers in a river sub-basin?

To answer these research questions, I sampled fishes by using traditional fishing methods (castnet) of different mesh sizes in four river sub-basins. A stream segment was the basic sampling unit. Sampled fishes were identified in the field and released back into the water soon after taking their body measurements (total length in cm). Standard textbook and identification keys were followed (Jayaram, 2010, Daniels, 2002) and experts consulted for species level identification. Systematic and rigorous field sampling resulted in recording of 93 fish species belonging to nine orders and 18 families with 18322 individuals. Malaprabha was the species rich sub-basin (53 species) followed by Mhadei (47 species), Tunga (45 species) and Bhadra (24 species). The family Cyprinidae was dominant in all the sub-basins depicting common pattern found in South Asian rivers (Bhat, 2003; 2004). This study also resulted in the discovery of a new fish species – Kudremukh barb (Pethia striata) from the headwater regions of Tunga basin (Atkore et al., 2015) and reported a healthy population of critically endangered fish i.e. Wayanaad mahseer (Barbodes wyanaadensis) in tributaries of Bhadra and Tunga in the Kudremukh National Park. Hollow shape of the rank abundance curve suggested most species in these sub-basins are rare and few species are abundant (Magurran, 2004). At the scale of a segment- water chemistry, and at the larger spatial scales -stream order determined native fish diversity (richness and abundance) in four river sub-basins. Higher fish species richness in segments was associated with higher disturbance levels while higher abundance was associated with least disturbance. This indicates occurrence of fish recruitment (fries and fingerlings) in relatively less disturbed segments.

This study is one of the first to examine the species turnover both within and across

viii river basins. Fish turnover was governed by stream order, stream substrate composition, canopy cover along the stream banks and water quality variables in the two basins situated within the protected area boundary. Thus, maintaining free- flowing river stretches takes on greater importance which allows persistence of diversity of stream habitats and native fish species composition. At the basin wide scale assessing the functional characteristics such as guilds of fishes and their responses to environment and various anthropogenic threats is an emerging and critical area of research in the freshwater fish community ecology (Winemiller, 2010).

Previous studies have assessed fish guild structure in a few river systems in India (Daniels, 2002; Bhat, 2003; Johnson and Arunachalam, 2010; Chakrabarty and Homechaudhuri, 2013; Kundu et al., 2014). These studies have generated important insights on fish guilds but, none of these studies have quantified either the relationship between various fish guilds and stream characteristics including water quality variables and hydrological barriers. I attempted to fill this important knowledge gap by evaluating the responses of fish guild richness to hydrological barriers and water quality variables across four river basins. The results indicated that the water column based positional fish guilds responded to diverse water chemistry variables highlighting the complexity involved in the fish-environment relationship. However, certain guilds especially surface-dwelling guilds were negatively affected due to hydrological barriers while mid-column and bottom dwelling guilds benefited from the impoundment effect (Aarts and Nienhuis, 2003; Kanno et al., 2010). This study highlights the importance of water chemistry monitoring in river basins and raises concerns over the effect of hydrological barriers on fish guild structure. This study also indicates the importance of incorporating long-term data on stream discharge and water chemistry to determine guild specific responses to either modified flow regime or changing water chemistry (Macnaughton et al., 2016).

This research also is one of the first in India and Southeast Asia to demonstrate the evidence of fish species recovery downstream of hydrological barriers in sub-basins due to the contribution of undammed tributaries and stream environmental variables. The estimated species recovery (proportion of species found in the undisturbed/control segment to that of disturbed segments) downstream of a dam was 90% at distances of 2 and 5 km. This attributed to the contribution received from the

ix undammed tributaries. These undammed tributaries have ameliorated the water chemistry condition in the main river channel and also served as a refuge to many endemic or rare stream fishes. Based on my study catchments varying between 160 to 1332 sq.km and following the precautionary principles for conserving highly endangered biodiversity, I recommend maintaining a minimum of 5-10 undammed or undisturbed streams in a 500 sq.km sub-basin to mitigate the effect of existing or planned hydrologic barriers.

In a nutshell, this study demonstrates and highlights the importance of comprehensive field sampling effort covering multiple spatial scales within and across river basins. River basin managers could effectively use data generated from this study to formulate effective scale-dependent river conservation guidelines and monitor freshwater resources in the future.

Table of contents

Sl No Title Page No.

1 Acknowledgements i-v

2 Executive Summary vii-x

3 Chapter 1: Introduction 1

4 Chapter 2: Patterns of fish diversity across river basins in the Western Ghats hotspot 13

5 Chapter 3: Determinants of native fish species turnover in the Western Ghats, India 49

6 Chapter 4: River barriers and environmental variables influence fish guild responses in the Western Ghats of India 71

7 Chapter 5: Assessing the recovery of fish assemblages downstream of hydrological barriers in India’s Western Ghats 100

8 Chapter 6: Synthesis and Conclusion 123

9 Appendix 2.1a: Details of sampling locations and disturbance regimes in four sub-basins 13 6 2.1b: Systematic list of freshwater fish found in the study area 141 2.2: Range (minimum, maximum), mean and standard deviation of 143 environmental variables 2.3: Exploratory plots 144 2.4: Diagnostic plots 152 10 3.1 :Freshwater fish species encountered in both Bhadra and Tunga sub-basins in the Kudremukh National Park 153 3.2a: Species dissimilarity with respect to combined basin reference maximum across Bhadra and Tunga and b) species dissimilarity with respect to basin reference maximum for endemic species using incidence data 155 3.3a: Species dissimilarity with respect to the indiv- dual basin reference maximum using incidence data for all species and b) for endemic species. 15 6 Species dissimilarity is consistently higher in Tunga

than Bhadra basin. 11 4.1: Supporting information on catchment details across four sub-basins 157 Details of environmental variables measured 157 :Supporting information – S2 Table 1 15 8 :Table 2 on literature review for guilds sampled in the study area 163 :Supporting information – S3 Figure 1 165 :PCA loadings 166 5.1 Copeia paper 167 5.2 River Research and Application 168

xii

List of Tables

Table No Table title Page No.

2.1 a) Details of sampling locations in each of the river 20 basin. b) Details of catchment area, direction of flow and the 23 state in which the study is conducted. c) Sampling duration in each of the river sub-basin. 24 d) Fish richness with mean, SD and SE across stream order 29 e) Fish richness with mean, SD and SE across stream order 29 2.2 Parameter estimates for fish species richness and fish species abundance. 32 3.1 Mantel’s test shows the relationship between pairwise dissimilarity in species composition and dissimilarity in ecological covariates for Bhadra Tunga sub-basin (combined). 58 3.2 Mantel’s test shows the relationship between pairwise dissimilarity in species composition and dissimilarity in ecological covariates in Bhadra sub-basin. Species turnover was driven mainly by stream characteristics (stream order, canopy cover) and local habitat effects (rocks). 58 3.3 Mantel’s test shows the relationship between pairwise dissimilarity in species composition and dissimilarity in ecological covariates in Tunga sub-basin. 59 4.1 The Mann-Kendall (M-K) test of significance for total rainfall (mm/yr), and maximum and minimum discharge (m3/s). A sharp declining monotonic trend in river discharge is noted from 1979 to 2013 despite no significant trends in rainfall time-series, which is likely due to river flow regulation and abstraction of water for irrigation. 82 4.2 Comparison of environmental variable values in NRSB and RSB. Statistical significance (***) was obtained for differences in means using a Welch t-test (with unequal variance for groups) [alpha=0.05; p < 0.001***, ^ indicates no significant difference]. 86 4.3 Selected Generalized Linear Models (GLMs with Poisson errors) showing effect sizes of environmental variables on guild-wise species richness in all sub- basins, and for RSB and NRSB. Delta AIC values indicate improvement brought in fit by GLM over null (intercept-only) model. In case of MCD species richness in NRSB, no model was found to improve fit over the null model. 88 5.1 Site-wise details on locations of barriers, cumulative

xiii

impacts and number of undammed tributaries present between segments and their nearest upstream barriers. 105 5.2 Species sampled in the study area, occurrence up to 2 and 5 km distances downstream of barriers, with life- history traits and recovery levels. 111 5.3 Non-linear regressions used to model fish recovery as a function of downstream distance from barriers. Total and endemic fish recovery was characterized as rising

immediately after a minimum distance threshold, as

predicted consistently by the asymptotic exponential models. The chosen models are highlighted in bold. 113

xiv

List of Figures

Figure No Figure legend Page No.

2. 1 a) Sampling locations in the four river sub-basins of the Western Ghats, India. (b) Malaprabha-Mhadei (c) Tunga-Bhadra 18 d) Schematic diagram showing a nested sampling design. Figure on left indicates relatively undisturbed river systems and figure on the right shows relatively more disturbed river system. 19 2. 2 a) The species accumulation curve for each river sub- basin b) species accumulation across number of individuals, c) The species accumulation curve for dry and wet season using number of segments sampled and d) pooled individuals from all the sampled segments 28 2.3 a) Dominant fish families b) species abundance rank distribution showing five dominant fish species across four river sub-basins. 28 2. 4 a) The relationship between fish species richness and stream orders b) species abundance and stream order. 29 2. 5 Endemic species richness in different stream orders. 30 2. 6 Fish species richness and habitat types (a, b). 34 2. 7 a) Species richness and disturbance level b) Species abundance and disturbance level 35 3. 1 Sampling locations in both Tunga and Bhadra river sub-basins. 52 3. 2 Relationship between species dissimilarity as a function of geographical distance for Bhadra Tunga combined. Correlogram shows a sine-hole pattern where sites which are closer are also geographically similar (indicated by filled circles) than sites that are farther apart. Species turnover is high up to a distance of 15 km and strongly correlated with geographical distance (Mantel’s r = 0.16, p=0.0001) and the effect of species turnover becomes positive after a distance of 10 km distance. 55 3.3 Correlogram showing the relationship between pairwise dissimilarity and geographical distance in Bhadra. Species turnover effect is weakly correlated with geographic distance (r =-0.013, p = 0.97) and it becomes positive after 5 km distance in Bhadra. 56 3.4 Correlogram showing the relationship between pairwise dissimilarity and geographical distance in Tunga. Compositional dissimilarity (species turnover) showed a sine-hole pattern with high species turnover at smaller spatial scales (up to a distance of 1-2 km). In other words, compositional

dissimilarity was strongly associated with geographical distance (Mantel r=-0.21, p =0.002) 57 4.1 Sampling locations in the four sub-basins in the central Western Ghats, India. 76 4.2 Time-series of maximum (left column) and minimum (right column) values of discharge (cumecs, i.e. m3/s) and total rainfall (mm/yr) for three hydrological gauging stations in the Mhadei and Malaprabha sub- basin. 81 4.3 NMDS ordination axes and plot shows three clusters showing correlated guild classifications for fish species across the four sub-basins. 83 4.4 NMDS ordination plot shows fish species composition across the four sub-basins (open circles indicate species). Non-regulated (Bhadra, Tunga) and regulated basins (Malaprabha, Mhadei) differed distinctly in species composition. The Mhadei and Malaprabha also clearly differed from each other in species composition. 84 4.5 Guild-wise fish species richness in non-regulated and regulated sub-basins. Surface-dwelling guild richness (a) was higher in non-regulated sub-basin, while mid- column (b) and bottom-dwelling richness (c) was lower in non-regulated as compared to regulated sub- basins. 85 4.6 Differences in selected water quality variables between regulated and non-regulated river sub-basins 86 5.1 (a) Schematic showing the application of the Serial Discontinuity Concept (SDC) in arriving at the hypotheses tested in the study (b) Expected pattern of change in species. 104 5.2 Map showing fish species recovery (shown for total fish species) in the Malaprabha basin. Locations of upstream control segments (unregulated reference sites) and segments downstream of barrier (test segments) are shown in the inset 105 5.3 Increase in (a) total and (b) endemic fish species recovery with increasing distance from upstream barriers; reduction in recovery (c) at higher levels of cumulative impact (number of upstream barriers/ downstream distance in km) and improvement in species recovery with the number of undammed tributaries joining the river below barriers (d). Dissolved oxygen (DO) (e) increased with distance from barrier, and total alkalinity (f) reduced with distance from barrier. 110

xvi

List of Abbreviations

AIC Akaike Information Criteria APHA American Public Health Association APHRODITE Asian Precipitation- Highly Resolved Observational Data Integration Towards Evaluation BD Bottom dwelling GLM Generalized Linear Models IUCN International Union for Conservation of Nature KNP Kudremukh National Park MCM Million Cubic meters MCD Mid-column dwelling MK test Mann- Kendall’s test of significance NRSB Non-regulated sub-basin NMDS Non-metric Multi-dimensional Scaling PCA Principal Component Analysis RCC River Continuum Concept RSB Regulated sub-basin SAC Species Accumulation Curve SD Surface dwelling SDC Serial Discontinuity Concept VIF Variation Inflation Factor WG Western Ghats WRDO Water Resources Development Organization ZSI Zoological Survey of India

xvii

Chapter 1

Introduction

Freshwater ecosystems and associated habitats harbor incredible biodiversity and support human well-being. Rivers, integral part of freshwater ecosystems, offer natural flow regimes that maintain hydrological, biogeochemical and ecological connectivity between channels, floodplains, wetlands and estuaries (FIU-GLOWS, 2016). Altered or modified flow regimes disrupt the natural processes that maintain biodiversity and ecosystem function (Bunn and Arthington, 2002; Poff, 1997). River systems are under constant pressure due to increasing societal demand for both energy and water, and as a result they are heavily exploited worldwide (Winemiller et al., 2016).

Owing to such increasing threats, river systems across the globe are receiving growing conservation attention. Existing knowledge of terrestrial ecosystems are the prime driving force for aquatic ecologists to understand whether the same factors that drive species assemblage patterns in terrestrial ecosystems also drive patterns in aquatic ecosystems. Past studies conducted in temperate regions on wetland (lakes and estuaries), streams and rivers suggest that diversity patterns are largely influenced by variety of factors depending on scale (Gido and Jackson, 2010; Matthews, 1998; Winemiller et al., 2008). For instance, both biotic (competition, predation) and abiotic factors (catchment characteristics, stream channel morphology, habitat heterogeneity, water chemistry, changes in land use pattern and disturbance regime) influence species richness patterns at local, regional, basin or continental scales. In lake ecosystems, depth is a crucial factor that defines species richness patterns, while in rivers and streams, the natural flow regime is found to be the important variable maintaining habitat as well as stream biota (Bunn and Arthington, 2002; Poff, 1997; Poff and Allan, 1995).

This idea was recognized by temperate river ecologists who formulated the River Continuum Concept (RCC) (Vannote et al., 1980), which states that stream biota are shaped by energy flowing from the headstream to downstream areas. This idea has generated wide interest among river ecologists globally. Since then, numerous studies

1 tested and replicated this both in temperate as well as tropical regions. This concept only considers river/stream as a single unit and does not take into account lateral connectivity. The Floodplain Concept emerged as a criticism to the RCC idea. Since river systems have been appropriated world-wide by dams and barrages, many scientists believed that the modified flow regime of the rivers has affected the river- dependent biota. These river barriers have fragmented riverine habitat into single or isolated habitats which gave rise to The Serial Discontinuity Concept (Ellis and Jones, 2016; Stanford and Ward, 2001; Stanford et al., 1996). All these concepts consider rivers as isolated entities and do not acknowledge the contribution from the terrestrial ecosystems in which they flow, and do not consider the fact that they are laterally connected ecosystems. This important gap was recently bridged by the idea of the Riverine Ecosystem Synthesis (Thorp et al., 2008) which consider rivers as integrated laterally connected systems. This holistic approach to studying rivers has been widely applied across the globe, especially focusing on key taxa such as macroinvertebrates and fishes since these groups are considered as biological indicators.

Studies focusing on stream/river ecosystems till date suggest that riverine biodiversity (species richness, abundance, species turnover and composition) is influenced by various factors either independently or in combination. The importance of these factors in influencing biodiversity is scale dependent. Global analyses on fish richness suggest that river size (surface area of the drainage basin and mean annual river discharge) and to a lesser extent, energy availability (net primary productivity), are important factors influencing fish species richness patterns (Oberdorff et al., 1995; 2011). At local scales, physical factors appear to determine species richness in variable environments, biological ones being more important under stable environmental conditions (Oberdorff et al., 1995; 2011). At regional and geographical scales, physical factors such as river size and climate along with historical factors such as speciation rates and dispersal are major determinants of species richness (Oberdorff et al., 1995; 2011). Patterns and processes observed in local assemblages are determined not only by local mechanisms acting within assemblages but also result from processes operating at larger spatial and temporal scales (Oberdorff et al., 1995; 2011). However, these patterns and processes are also influenced by anthropogenic impacts such as hydropower projects, river linking projects, deforestation, pollution and introduction of invasive species. Such increasing

2 anthropogenic impacts on river systems have imperiled a significant proportion of freshwater biodiversity leading to freshwater species extinction (Chakona and Swartz, 2012; Ricciardi and Rasmussen, 1999). A recent estimate based on modelled richness and river drainage relationships under climate change scenarios suggest that fish extinction may vary from 4-22% and might accelerate further under severe water scarcity (Tedesco et al., 2013).

Tropical and sub-tropical river systems have been extensively studied in terms of their fish biodiversity in relation to various anthropogenic impacts (Bailly et al., 2016; Dudgeon, 2000; 2006; Pandit and Grumbine, 2012; Unmack, 2001; Winemiller et al., 2008). In the past, limnological studies have assessed the relationship between water quality variables and fish diversity in lakes as well as in stream environments (Chapman, 1996; Lowe-Mc-Connell, 1975; 1987; Merz, 2013; Mattos et al., 2014). Studies demonstrated that habitat heterogeneity, stream characteristics and elevation influence fish species richness within as well as across river systems (Gilliam et al., 1993; Piet, 1998; Arunachalam, 2000; Araujo et al., 2008; Mendonca, 2005; Phomikong et al., 2015). Recently, numerous studies have examined the impact of dams on fish communities and have found that dams disconnect river habitats which influences fish community structure based on functional traits (Arthington, 2004; Arthington et al., 2014; Bhat and Magurran, 2007; Chakrabarty and Homechaudhuri, 2013; Vasconcelos et al., 2014a, b). However, few studies evaluate species composition and recovery downstream of dams (Storey et al., 1991; Jeffree et al., 2001).

Past studies on fish community structure in the Western Ghats region in India Although studies on patterns of fish diversity, distribution and conservation are on the rise, there are still large gaps in knowledge regarding fish diversity from the Western Ghats (WG) biodiversity hotspot in India. Previous studies have quantified the relationship between fish diversity and associated abiotic factors at local and regional scales (Bhat, 2002; Johnson and Arunachalam, 2009; 2010; Raghavan et al., 2008). In the southern WG, fish abundance was significantly correlated with stream habitat characteristics, mainly area and volume (Johnson and Arunachalam, 2010). Studies

3 also identified habitat utilization patterns of fishes and found five habitat use guilds. These studies were able to reconstruct a clear pattern of fish segregation in east and west flowing rivers of this region (Johnson and Arunachalam, 2009). In another study, fish species richness increased with increasing stream order and decreased with increasing elevation (Raghavan et al., 2008). This study also found a clear temporal pattern where fish species richness was highest during the day and lowest at dusk.

In the central WG, high fish species richness was recorded in four river systems (Bhat, 2005; Bhat and Magurran, 2007). Fish in these rivers were also found to show ecological and morphological partitioning. When the impact of disturbance on fish communities was examined, it was found that natural and unimpacted rivers (Aghnashini and Bedthi) had more structured species distribution than rivers with disturbance (Sharavati and Kali). Most of these studies suffer from several shortcomings; in particular, they did not consider river-basin-wide approach and were either focused on one or few river systems. Many of these studies also lack comprehensive data on water chemistry, and thus are unable to link how fish diversity is influenced by certain crucial water chemistry variables. For most of these river systems, there is no information regarding the relationship between disturbance regimes and functional traits of fish or quantification of species recovery in rivers with dams.

To fill this important knowledge gap in our understanding of the WG fish assemblage, I studied the drivers of native fish diversity at multiple spatial scales both within as well as across four river sub-basins of the WG region, in India. A basin-wide approach is known to offer a unique opportunity to study complete fish community organization at multiple spatial scales.

Study river basins In this study, I focused on four river sub-basins - Mhadei, Malaprabha, Tunga and Bhadra. The detailed description of each sub-basin can be found in Chapter 2 and 3. Of the four sub-basins, Mhadei is the only west flowing sub-basin that originates within the hilly region of Bhimgad Wildlife Sanctuary at Panshet cha nala at 760 m in Belgaum district of Karnataka state. The river flows through North Goa and finally

4 meets the estuary at Panaji in the state of Goa (Ibrahampurkar, 2012). Malaprabha, an east flowing tributary of Krishna river also originates in Belgaum district near Kankumbi hills at 760 m elevation. It flows through Belgaum and Dharwad districts before draining into Renuka Sagara reservoir at Soundatti, situated downstream, at a distance of 300 km (Atkore et al., 2012). Both these river basins are exposed to a variety of anthropogenic disturbances such as check dams, barrages, inter-basin water transfer canals, water uptake for irrigation (sugarcane fields, oilseed and rubber- oilpalm plantation), pollution, substrate mining, illegal fishing and vehicle washing among other disturbances.

Bhadra and Tunga are the two other rivers originating in the WG region of Kudremukh National Park, at an elevation 1160 m. Both these rivers flow in opposite directions. Bhadra drains southwards into the Kudremukh range while Tunga drains northwards into Kerekatte range forming Bhadra and Tunga sub-basins respectively. There are numerous perennial streams joining each river, giving rise to a diverse range of habitats such as cascades, pools, runs and riffles. The topography in Bhadra basin is steeper than in Tunga basin and there are barely any major hydrological barriers on these rivers except a few natural waterfalls.

Objectives of the study This study attempted to understand the drivers of native fish diversity at multiple spatial scales both within as well as across river basin. I digitized and delineated the stream network in each basin using GIS tools. The sampling in each basin was spread across different gradients viz. head stream to lower streams, different stream orders, and across habitat types to cover multi-spatial extent. A stream segment of 150 m length was determined as the basic sampling unit for the study which was nested in one, two or more habitat types within a segment. A segment was nested within a stream, a stream was nested within a sub-basin and a sub-basin was nested within an eco-region. At each segment, I collected data on stream characteristics as well as water chemistry variables. Details of study design can be found in Chapter 2, 3, 4 and 5.

In the second chapter, I examine the relationship between predictor variables such as stream characteristics and water chemistry variables with response variables such as species richness and abundance across sampled stream segments. I used Principal Component Analysis (PCA) to determine key variables that influenced the fish diversity. Then I used Generalized Linear Models (GLM) to uncover the relationship between fish richness and abundance and environmental variables including disturbance regime.

In the third chapter, I analyzed the drivers of species turnover in two adjacent river basins i.e. Bhadra (670 sq. km) and Tunga (328.60 sq. km) which were similar in their geography. I used Mantel’s correlogram and Bray-Curtis dissimilarity index (species- presence absence plus abundance data) to understand the spatial autocorrelation between and across sampled segments at various distances within each sub-basin. I compared my results with similar studies carried out elsewhere from tropical river systems.

The fourth chapter evaluates the relationship between hydrological barriers (check dam, barrages), and local environmental variables with fish guild richness in similar stream orders, elevation and habitat types in four sub-basins. Two analytical techniques used were Non-metrical multi-dimensional scaling (NMDS) approach (uses species abundance data to build a Bray-Curtis dissimilarity matrix) and GLM. First, I classified all the sampled fish into various fish guilds following standard literature (Welcomme, 1985; De Silva et al., 1979; Aarts and Nienhuis, 2003; Winemiller et al., 2008; Chakrabarty and Homechaudhuri, 2013; Kundu et al, 2014). NMDS was used to investigate whether the river basins differ in their guild species composition. I then examined what factors drive guild species richness across studied basins using GLMs. Box and whiskers plots were used to determine the relationship between fish richness and abundance with water quality and environmental variables.

Finally, I evaluated species recovery patterns of stream fish assemblages in one river basin (Malaprabha) to test the hypothesis that, the species recovery will improve downstream of a dam due to the contribution of undammed tributaries joining the main river channel and that recovery improves with increasing distance from dam. To test these hypotheses, I first calculated species recovery i.e. proportion of species

6 found downstream of dams, out of total and endemic species richness recorded in the upstream control segments unaffected by the dam. I used two types of GLM’s (logistic model and asymptotic exponential model) that reflected different predictions about how species recovery was expected to increase. I also calculated the cumulative impact of dam by measuring actual distance from each of the dams using GIS. Spearman’s correlation was used to find the relationship between the recovery (total and endemic species recovery) with various environmental and water quality variables.

I conclude with a synthesis of the findings from each research question investigated and further discuss the conservation implications of this study for WG rivers.

References

Aarts, Bram G.W., and Nienhuis, Piet H. 2003. Fish zonations and guilds as the basis for assessment of ecological integrity of large rivers. Hydrobiologia 500(1), 157-178. doi: 10.1023/a:1024638726162 Araújo, Francisco Gerson, Pinto, Benjamin Carvalho Teixeira, and Teixeira, Tatiana Pires. 2008. Longitudinal patterns of fish assemblages in a large tropical river in southeastern Brazil: evaluating environmental influences and some concepts in river ecology. Hydrobiologia 618 (1), 89. doi: 10.1007/s10750-008-9551-5 Arthington A H., K Lorenzen, B. J. Pusey, R. Abell, A. Halls, O. K. Winemiller, D A Arington, and E. Baran 2004. River fisheries: ecological basis for management and conservation. In: Proceedings of the Second International Symposium on the Management of Large Rivers for Fisheries Volume I. 31- 60. Arthington, A. H., Rolls, R. J., Sternberg, D., Mackay, S. J., and James, C. S. 2014. Fish assemblages in subtropical rivers: low-flow hydrology dominates hydro- ecological relationships. Hydrological Sciences Journal 59(3-4), 594-604. doi: 10.1080/02626667.2013.844345 Arunachalam, M. 2000. Assemblage structure of stream fishes in the Western Ghats (India). Hydrobiologia 430(1-3), 1-31. doi: 10.1023/A:1004080829388 Atkore, V., J. Krishnaswamy., S. Badiger., and Shanker, K. 2012. Preliminary assessment of fish fauna in the upper Malaprabha river basin of the Western Ghats, India. Paper presented at the 4th Bio-Asia Regional Seminar, Bio-Asia Program, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China. Bailly, Dayani, Cassemiro, Fernanda A. S., Winemiller, Kirk O., Diniz-Filho, José Alexandre F., and Agostinho, Angelo A. 2016. Diversity gradients of Neotropical freshwater fish: evidence of multiple underlying factors in human- modified systems. Journal of Biogeography 43(8), 1679-1689. doi: 10.1111/jbi.12749 Bhat, A. 2005. Ecomorphological correlates in tropical stream fishes of southern India. Environmental Biology of Fishes 73, 211-225. doi: 10.1007/s10641- 005-0561-0 Bhat, A. 2002. A study of the diversity and ecology of freshwater ﬁshes of four river systems of the Uttara Kannada District, Karnataka, India. A PhD thesis submitted to the Centre for Ecolgical Sciences, Indian Institute of Sciences, Bangalore.Pp. 178. Bhat, A., and Magurran, A. E. 2007. Does disturbance affect the structure of tropical fish assemblages? A test using null models. Journal of Fish Biology 70(2), 623-629. doi: 10.1111/j.1095-8649.2006.01252.x

Bunn, E. Stuart, and Arthington, H. Angela. 2002. Basic Principles and Ecological Consequences of Altered Flow Regimes for Aquatic Biodiversity. Environmental Management 30(4), 492-507. doi: 10.1007/s00267-002-2737-0 Chakona, Albert, and Swartz, Ernst R. 2012. Contrasting habitat associations of imperilled endemic stream fishes from a global biodiversity hot spot. BMC Ecology 12(1), 19. doi: 10.1186/1472-6785-12-19 Chakrabarty, M., and Homechaudhuri, S. 2013. Fish guild structure along a longitudinally-determined ecological zonation of Teesta, an eastern Himalayan river in West Bengal, India. Arxius de Miscel-lania Zoologica 11, 196-213. Chapman, D. 1996. Water quality assessment - a guide to the use of biota, sediments and water in environmental monitoring (Ed). Vol. Second: Chapman and Hall on behalf of UNESCO, WHO and UNEP. Pp. 653. De Silva, S.S., Cumaranatunga, P.R.T., and De Silva, C.D. 1979. Food, Feeding Ecology and Morphological Features Associated With Feeding of Four Co- Occurring Cyprinids (Pisces: Cyprinidae). Netherlands Journal of Zoology 30(1), 54-73. doi: doi:10.1163/002829680X00032 Dudgeon, David. 2000. The Ecology of Tropical Asian Rivers and Streams in Relation to Biodiversity Conservation. Annual Review of Ecology and Systematics 31, 239-263. Dudgeon, David, Arthington, Angela H., Gessner, Mark O., Kawabata, Zen-Ichiro, Knowler, Duncan J., Lévêque, Christian, and Sullivan, Caroline A. 2006. Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews 81(02), 163-182. doi: doi:10.1017/S1464793105006950 Ellis, L. E., and Jones, N. E. 2016. A Test of the Serial Discontinuity Concept: Longitudinal Trends of Benthic Invertebrates in Regulated and Natural Rivers of Northern Canada. River Research and Applications 32(3), 462-472. doi: 10.1002/rra.2861 FIU-GLOWS. 2016. Freshwater inflow requirements for Wami River Estuary, Sadani National Park, Tanzania. ISBN 978-1-941993-14-9. Pp. 104. Gido, K., and Jackson, Donald A. 2010. Community Ecology of Stream Fishes: Concepts, Approaches, and Techniques. Bethseda, Maryland, USA: American Fisheries Society Sympposium 73. Pp. 665. Gilliam, James F., Fraser, Douglas F., and Alkins-Koo, Mary. 1993. Structure of a Tropical Stream Fish Community: A Role for Biotic Interactions. Ecology 74(6), 1856-1870. doi: 10.2307/1939943 Ibrahampurkar, M. M. 2012. Hydrological and hydrogeological evaluation of Mhadei river watershed in Goa and Karnataka. A Ph.D thesis submitted to The Department of Earth Science. Goa University, Panjim. Pp 282. Jeffree, Ross A., Twining, John R., and Thomson, Jeremy. 2001. Recovery of Fish Communities in the Finniss River, Northern Australia, Following Remediation

of the Rum Jungle Uranium/Copper Mine Site. Environmental Science and Technology 35(14), 2932-2941. doi: 10.1021/es001719+ Johnson, J.A. , and Arunachalam, M. 2009. Diversity, distribution and assemblage structure of fishes in streams of southern Western Ghats, India. Journal of Threatened Taxa 1(10), 507-513. Johnson, Jeyaraj Antony, and Arunachalam, Muthukumarasamy. 2010. Habitat use of fishes in streams of Kalakad Mundanthurai Tiger Reserve, India. International Journal of Ecology and Development 17 (F10), 34-47. Kundu, N., D. Mazumdar, and Homechaudhuri, S. 2014. Seasonal changes in fish assemblages and trophic guild in selected interconnected rivers in eastern India. Indian journal of Fisheries 61(3), 35-44. Lowe-McConnell, R. H. 1975. Fish communities in tropical freshwaters their distribution, ecology and evolution.: London.Pp 337. Lowe-McConnell, R. H. 1987. Ecological studies in tropical fish communities. Cambridge: Cambridge University Press, Cambridge.Pp. 382. Matthews, M. J. 1998. Patterns in freshwater fish ecology. Massachusetts, USA.: Kluwer Academic publisher.Pp. 670. Mattos, Taylan M., Costa, Marcus R., Pinto, B. C. T., Borges, Joyce L. C., and Araújo, F. Gerson. 2014. To what extent are the fish compositions of a regulated river related to physico-chemical variables and habitat structure? Environmental Biology of Fishes 97(6), 717-730. doi: 10.1007/s10641-013- 0175-x Mendonca, F.P., W.E. Magnusson and J. Zuanon. 2005. Relationship between habitat characteristics and fish assemblages in small streams of Central Amazonia. Copeia 4, 751-764. Merz, Sinclair Knight. 2013. Characterizing the relationship between water quality and water quantity (Environment Department of Sustainability, Water, Population and Communities, Trans.) St Leonards NSW 2065 Australia. Pp. 139. N. LeRoy Poff, J. David Allan, Mark B. Bain, James R. Karr, Karen L. Prestegaard, Brian D. Richter, Richard E. Sparks, and Julie C. Stromberg 1997. The natural flow regime: a paradigm for river conservation and restoration. BioScience 47(11 ), 769-784 doi: 10.2307/1313099 Oberdorff, Thierry, Gu, xe, gan, Jean-Fran, xe, ois, and Hugueny, Bernard. 1995. Global Scale Patterns of Fish Species Richness in Rivers. Ecography 18(4), 345-352. Oberdorff, Thierry, Tedesco, Pablo A., Hugueny, Bernard, Leprieur, Fabien, Beauchard, Olivier, Brosse, Sébastien, and Dürr, Hans H. 2011. Global and Regional Patterns in Riverine Fish Species Richness: A Review. International Journal of Ecology. doi: 10.1155/2011/967631

Pandit, Maharaj K., and Grumbine, R. Edward. 2012. Potential Effects of Ongoing and Proposed Hydropower Development on Terrestrial Biological Diversity in the Indian Himalaya. Conservation Biology 26(6), 1061-1071. doi: 10.1111/j.1523-1739.2012.01918.x Phomikong, P., Fukushima, M., Sricharoendham, B., Nohara, S., and Jutagate, T. 2015. Diversity and Community Structure of Fishes in the Regulated Versus Unregulated Tributaries of the Mekong River. River Research and Applications 31(10), 1262-1275. doi: 10.1002/rra.2816 Piet, GerbrandJ. 1998. Ecomorphology of a size-structured tropical freshwater fish community. Environmental Biology of Fishes 51(1), 67-86. doi: 10.1023/A:1007338532482 Poff, N. LeRoy, and Allan, J. David. 1995. Functional Organization of Stream Fish Assemblages in Relation to Hydrological Variability. Ecology 76(2), 606-627. doi: 10.2307/1941217 Poff, N.L. 1997. Landscape filters and species traits: towards mechanistic understanding and prediction in stream ecology. Journal of the North American Benthological Society16:391-409.

Raghavan, Rajeev, Prasad, G., Ali, P. H. Anvar, and Pereira, Benno. 2008. Fish fauna of Chalakudy River, part of Western Ghats biodiversity hotspot, Kerala, India: patterns of distribution, threats and conservation needs. Biodiversity and Conservation 17(13), 3119-3131. doi: 10.1007/s10531-007-9293-0 Ricciardi, Anthony and Rasmussen, Joseph B. 1999. Extinction Rates of North American Freshwater Fauna. Conservation Biology 13(5), 1220-1222. doi: 10.1046/j.1523-1739.1999.98380.x Stanford, Jack A., and Ward, J. V. 2001. Revisiting the serial discontinuity concept. Regulated Rivers: Research and Management 17(4-5), 303-310. doi: 10.1002/rrr.659 Stanford, Jack A., Ward, J. V., Liss, William J., Frissell, Christopher A., Williams, Richard N., Lichatowich, James A., and Coutant, Charles C. 1996. A general protocol for restoration of regulated rivers. Regulated Rivers: Research and Management 12(4-5), 391-413. Storey, A. W., Edward, D. H., and Gazey, P. 1991. Recovery of aquatic macroinvertebrate assemblages downstream of the Canning Dam, Western Australia. Regulated Rivers: Research and Management, 6(3), 213-224. doi: 10.1002/rrr.3450060306 Tedesco, Pablo A., Oberdorff, Thierry, Cornu, Jean-François, Beauchard, Olivier, Brosse, Sébastien, Dürr, Hans H., Hugueny, Bernard. 2013. A scenario for impacts of water availability loss due to climate change on riverine fish extinction rates. Journal of Applied Ecology 50(5), 1105-1115. doi: 10.1111/1365-2664.12125

Thorp, James H., Martin C. Thoms, and Michael, D. Delong. 2008. The riverine ecosystem synthesis: Towards Conceptual Cohesiveness in River Sciece: Academic Press. Pp.223. Unmack, P T. 2001. Biogeography of Australian Freshwater fishes. Journal of Biogeography 28(9), 1053-1089. Vannote, R. L., Minshall, G. W., Cummins, K. W., Sedell, J. R., and Cushing, C. E. 1980. The River Continuum Concept. Canadian Journal of Fisheries and Aquatic Sciences 37(1), 130-137. doi: 10.1139/f80-017 Vasconcelos, L. P., Alves, D. C., and Gomes, L. C. 2014a. Fish reproductive guilds downstream of dams. Journal of Fish Biology 85(5), 1489-1506. doi: 10.1111/jfb.12501 Vasconcelos, L. P., Alves, D. C., and Gomes, L. C. 2014b. Spatial and temporal variations among fish with similar strategies: patterns of reproductive guilds in a floodplain. Hydrobiologia 726(1), 213-228. doi: 10.1007/s10750-013-1767-3 Ward, J. V., and Stanford, J. A. 1995. The serial discontinuity concept: Extending the model to floodplain rivers. Regulated Rivers: Research and Management, 10(2-4), 159-168. doi: 10.1002/rrr.3450100211 Welcomme, R. L. 1985. River Fisheries FAO Fisheries Technical Paper Pp. 330. Winemiller, K. O., McIntyre, P. B., Castello, L., Fluet-Chouinard, E., Giarrizzo, T., Nam, S.,. Sáenz, L. 2016. Balancing hydropower and biodiversity in the Amazon, Congo, and Mekong. Science 351(6269), 128-129. doi: 10.1126/science.aac7082 Winemiller, Kirk O., Agostinho, Angelo A., and Caramaschi, Érica Pellegrini. 2008. Fish Ecology in David Dudgeon (ed). Tropical Stream Ecology Pp. 107. London: Academic Press.

Chapter 2 Patterns of fish diversity across river basins in the Western Ghats hotspot

Publication Atkore, V. M., Knight, J.D. M., Rema Devi and J. Krishnaswamy. (2015). A new species of Pethia from the Western Ghats, India. Copeia. 103 (2): 290-296.

Patterns of fish diversity across river basins in the Western Ghats hotspot

Introduction Biodiversity is influenced by different factors across different spatial scales (DeFries et al., 2010; Gaston, 2000). Ecologists, over the last few decades, have increasingly focused on understanding the factors influencing biodiversity at different spatial scales, from the local to the global. It is often thought that local-scale diversity is influenced by species interactions (Jackson et al., 2001; Matthews, 1998) while at larger spatial scales, processes of diversification and dispersal (Oberdorff et al., 2011; Olden et al., 2010) influence diversity. There have been extensive studies on terrestrial ecosystems and the scaling of biodiversity (MacArthur, 1972; Rozenweig, 1995; van Jaarsveld et al., 1998; Schwartz et al., 1999; Poiani et al., 2000). However, aquatic and river systems can also be studied using a similar framework, and such studies are providing insights into the distribution and organization of aquatic communities across spatial scales. River systems, however, have some distinct characteristics as they are embedded within distinct topographically defined catchments, and this limits species dispersal both within (due to the stream flow from upstream to downstream and river barriers e.g. waterfalls) and between river basins. Fish diversity within river systems is also governed by different factors at different spatial and temporal scales (Hugueny et al., 2010; Oberdorff et al., 2011).

Fish diversity at local scales (stream length between 100-150 m segments) is influenced by abiotic factors, water chemistry (Matthews, 1998), habitat structure (Gorman and Karr, 1978), as well as biological interactions (Holomuzki et al., 2010; Jackson et al., 2001). At the regional or eco regional level (100 - 1000 sq.km) species interactions are driven by catchment area, stream flow (Jackson et al., 2001), land-use characteristics (Schlosser, 1991), pollution (Magurran and Dawn, 2001) hydrological regulation (Sakaris, 2013) and disturbance regime (Dornelas, 2010). Fish communities are also structured across temporal scales; at much longer i.e. historical or evolutionary time scales, they are shaped by speciation rates and dispersal (Olden et al., 2010; Oberdorff et al., 2011), while at much shorter temporal scales (months to years) they are influenced by seasonality, flow variability and factors such as

14 frequency and magnitude of flood events (Bunn and Arthington, 2002; Jardine et al., 2015). Over the last few decades, ideas from community ecology (i.e. species-area relationship, latitudinal gradient in diversity etc.) have been widely applied in tropical river systems especially in the Latin America (Araujo et al., 2008; Maltchik, 2010; Tondato, 2010), Australia (Pusey et al., 1998; Unmack, 2001), New Zealand (Astorga et al., 2014; Burridge et al., 2008) and in South East Asia (Edds, 1993; Wikramnayake, 1990; Yap, 2002) to understand patterns of diversity and its scaling. However, our understanding of diversity, distribution and habitat ecology of many threatened freshwater fish groups in India is incomplete especially across spatial scales (Bhat, 2003; Molur et al., 2011).

In regulated basins, river discharge has been identified as the main driver of fish diversity (Bunn and Arthington, 2002; Bhatt, 2012), but fish distribution is also influenced by water chemistry or the “quality of water” from the viewpoint of fish biology and ecology (Menni et al., 1996; Alabaster, 2013) which in turn influences both species as well as functional diversity (Menni et al., 1996; Brasher, 2003; Pool et al., 2010; Macnaughton et al., 2016). For instance, water temperature and conductivity regulate fish movement and abundances of certain fish species (Chapman, 1996; Abes and Agostinho, 2001) whereas stream flow modifies fish- environment relationships via water chemistry variables (Fialho et al., 2008). In addition, different fish guilds are affected by artificial or seasonal discharges from dams. For instance, rheophilic-surface dweller-sensitive fish guilds which require running water habitat (run/riffles) (Winemiller and Jepsen, 1998; Vokoun, 2009) are negatively affected due to reduction of water connectivity below dams and barrages and at the same time, eurytopic-bottom dwelling-generalist fish guilds may benefit due to impoundment effect of dams (Vokoun, 2009; Chakrabarty and Homechaudhuri, 2013; Macnaughton et al., 2016). Certain river segments, especially in hydrologically regulated river systems, are exposed to higher water temperature, changes in water quality (e.g. increase in the calcium hardness concentration, reduction in dissolved oxygen level and anthropogenic disturbance). The responses of diverse fish guilds are complex as their preferred water chemistry environments are largely unknown (Chea et al., 2016 a, b).

Studies from the Western Ghats (WG) region, a global biodiversity hotspot, have

15 documented diversity and distribution patterns associated with biotic and abiotic factors at both local to regional scales focusing on single or few river systems but not entire basins except one study on the Godavari basin (Khedkar et al., 2014). Past research has shown a clear pattern of segregation in fish assemblages in east and west flowing rivers of the WG region (Johnson and Arunachalam, 2009). It was demonstrated that fish abundance was significantly correlated with stream habitat characteristics, mainly habitat area and volume (Johnson and Arunachalam, 2010). In the central WG, research has shown the presence of high fish species richness in four river systems with ecological and morphological partitioning of fishes in these rivers (Bhat, 2003, 2004; Bhat and Magurran, 2007). It was also found that, more natural and less impacted rivers (Aghnashini and Bedti in WG) had higher species richness than rivers with higher disturbance and hydrological modification (Sharavati and Kali in WG). In the southern WG, fish species richness increased with increase in stream orders and decreased with increasing elevation (Raghavan et al., 2008) while endemic fish richness peaked at mid-elevations (Abraham and Kelkar, 2012). While these studies provide valuable information, basin-wide surveys provide a more holistic understanding of community organization at multiple spatial scales that is essential for fish conservation as a part of river basin management (Hauer and Lamberti, 2007; Olden et al., 2010; Tedesco et al., 2012). Many small-to-large hydroelectric power projects are being developed in this global biodiversity hotspot and the absence of baseline and site specific data on fish species richness, distribution, and life history traits severely constrains conservation planning. These data will not only enable in establishing linkages between fish communities and their environmental factors, but can also form a basis for guiding river management plans in future.

Against this background, I examined patterns of diversity and distribution of freshwater fishes in four river sub-basins in the central WG, India. The specific objectives were to: a) Examine patterns of fish species richness and abundance in four sub-basins b) Investigate environmental drivers of species richness and abundance in these rivers, and c) Determine the effect of disturbance on fish species richness. I hypothesized that water quality parameters, especially dissolved oxygen, electrical conductivity, etc would influence fish species richness (Merz, 2013; Sheldon and

Fellows, 2010; Mattos et al., 2014). Fish species richness is also expected to be higher in the least disturbed segments due to the habitat complexity and available food resource (Grenouillet et al., 2002) than more disturbed segments.

Materials and Methods Study river basins I studied fish diversity in four sub-basins in the WG namely, Bhadra, Tunga (both non-regulated), Malaprabha and Mhadei (both regulated) situated in the states of Karnataka and Goa (Fig. 2.1). Each of these river basins is discussed in detail below:

(1) Malaprabha originates near Kankumbi village in Belgaum district of Karnataka state at 790 m elevation (Figure 2.1a). Malaprabha is an east flowing river with few seasonal streams and is the principal tributary of river Krishna. The catchment area is largely dominated by tropical wet evergreen forest in the northern side and semi- evergreen forest in the southern side of the river basin (Malkhede, 2003). Over the last decade, the cropping pattern along the river has shifted from legume (pulse crop) and cotton to sugarcane (Heller et al., 2012) resulting in large abstraction of ground and surface water, leading to reduced streamflow especially in dry-season. The river is exposed to a range of disturbances from headstream region near Kankumbi to downstream region at Khanapur. These include inter-river basin irrigation canal, abstraction of water, construction of temporary embankments/impoundments, sand- and-boulder mining and solid waste and untreated domestic sewage disposal from towns. As a result of this, the environmental and ecological health of the river has deteriorated rapidly. Therefore I classified this river sub-basin as a regulated sub- basin. (2) Mhadei is a west flowing river that originates near Degaon at an elevation of 685 m in the WG of Karnataka state. The river traverses through North Goa and finally meets the Arabian Sea in Panaji, Goa. The headwater region has numerous streams and some of the principal tributaries include Kalasa and Bhanduri. The Karnataka government has a long-

Figure 2. 1. (a) Sampling locations in the four river sub-basins of the Western Ghats, India (b) Malaprabha-Mhadei (c) Tunga-Bhadra

Figure 2. 1 (d). Schematic diagram showing a nested sampling design. Figure on left indicates relatively undisturbed river systems and figure on right shows relatively more disturbed river system.

-standing demand to divert water from these tributaries via an inter-basin transfer project to the Malaprabha river. The northern part of the basin is largely dominated by moist deciduous and evergreen forest while the southern part consists largely of agriculture and plantation. Water quality is affected by effluent from domestic water uses and disposal of untreated sewage. In addition, water abstraction for irrigation, agriculture and other domestic use is affecting river health. I classified Mhadei as another regulated sub-basin.

The other two river basins, (3) Bhadra and (4) Tunga, originate at an elevation of 1160 m near in the WG Gangamoola hill range in the Kudremukh National Park in the Chikkamgaluru and Dakshin Kannada districts of Karnataka state. The park is one of the UNESCO world heritage sites in the WG, dominated by wet evergreen forest and shola grassland (Kasturirangan, 2013). Each river is fed by numerous small to medium-sized perennial streams. I identified nine major streams in Bhadra and ten major streams in Tunga sub-basin. Water quality is not severely affected by various anthropogenic pressure. Bhadra drains into Bhadra reservoir situated at 142 km from Gangamoola. Tunga joins Bhadra downstream, thereby becoming the Tungabhadra river. Subsequently, it joins river Krishna at Kudalsangama in the Bagalkot district of

Karnataka state and finally meets the Bay of Bengal. (Fig. 2.1). Detail of each river basin is given in Table 2.1 a-c.

Sampling strategy I delineated the catchment area for each river sub-basin using Geographic Information System tools (ArcMap 10.1, Table 2.1a and b). I followed nested sampling design wherein each sample consisted of a river segment of approximately 100-150 m length. A segment consisted of a set of one or multiple channel units or meso-habitats including run, pools, runs, riffles and a cascade or a sequence of pool-run, run-riffle and riffle-pool. These segments were situated in different stream orders (1-7) in a sub- basin. I sampled such segments from the headstream to downstream (Table 2.1a, Fig. 2.1d). Systematic sampling was undertaken during different seasons. The total number of samples as per the nested sampling design in different seasons are as shown in Table 2.1c.

Table 2. 1a details of sampling locations in each of the river basin. River basin/Streams Seg. name Elevation (m) Stream order No of seg.

(1) Bhadra sub-basin Bhadra Kadambi joint 829 3 13 Bh7 860 3 Oldgrass point 854 3 Kurinjikal 830 3 Khasigadde 834 3 Bhagvati 825 4 Pandarmakki road 824 4 Bh-kachige joint 778 4 Khagundi 761 5 KIOCL 770 5 Nagraj mane 760 5 Kudremukh bridge 755 5 Hosmakki joint 750 5 Biligal Biligal1 775 4 03 Biligal3 805 3 Biligal5 816 3 Kachige Kachige1 778 4 04 Kachige3 785 4 Kachige4 790 4 Kachige5 806 4 Singsar Singsar1 810 4 03 Singsar2 814 4 Singsar3 928 4

River basin/Streams Seg. name Elevation (m) Stream order No of seg.

Kunya Kunya1 780 3 05 Kunya2 782 3 Kunya3 795 3 Kunya4 824 3 Kunya5 829 4 Carman Carman1 795 4 04 Carman2 810 4 Carman3 814 3 Carman4 898 3 Hosmakki Hosmakki1 910 1 03 Hosmakki2 920 4 Hosmakki3 930 4 Nelibedu Hosmakki4 767 3 03 Nelibedu1 802 3 Nelibedu3 809 3 Nelibedu5 861 3 Somvati Somvati1 1053 2 02 Somvati2 802 3 (2) Tunga sub-basin Mudba Mudba1 664 4 05 Mudba2 670 4 Mudba4 675 4 Mudba5 700 3 Mudba8 700 3 Mundsar Mundsar1 652 4 04 Mundsar2 667 4 Mundsar4 675 4 Mundsar8 700 4 Turad Turad1 642 4 04 Turad2 689 4 Turad3 690 4 Turad4 754 4 Karuchar Karuchar1 667 4 06 Karuchar2 672 4 Karuchar3 686 4 Karuchar4 691 4 Karuchar5 693 4 Karuchar6 697 4 Korkan Korkan1 656 4 06 Korkan2 661 4 Korkan3 675 4 Korkan4 676 4 Korkan6 679 4 Korkan11 880 4 Gangehole Ghole2 776 3 04 Ghole4 773 4 Ghole6 716 4 Ghole8 698 3

River basin/Streams Seg. name Elevation (m) Stream order No of seg.

Muje Muje3 662 3 03 Muje4 667 3 Muje7 680 4 Tanikod Tanikod2 736 2 02 Tanikod4 811 2 Vimala Mudba-mundsar joint 761 4 08 Tn2 681 5 Acharmakk 677 5 Tumbahalla 673 5 Before Acharmakki 670 5 Keshav mane 667 5 Before Kerekatte1 665 5 Before Kerekatte2 665 5 Tunga Kerekatte 656 5 11 Kerekatte school 655 5 Trogen point 650 5 Before Anand mane 650 5 Anand mane 650 5 Yadgar 650 5 Muje joint 645 6 Tanikod joint 645 6 Toursit point 635 6 Salmara 630 6 Nemmar 630 6 (3) Malaprabha sub-basin Haltar nala Shiroli 672 4 02 Shedegali 650 4 Mangetri nala Katgali 713 4 02 Valmiki 700 4 Malaprabha Amta 720 3 11 Torali 715 3 Devachihatti 720 3 Habbanhatti 714 3 Kusmali 700 3 Malavi 711 3 Olmani 693 3 Shankerpeti 653 4 Asoga 643 4 Rumewadi 655 5 Kupatgiri 650 5 (4) Mhadei sub-basin Mhadei Degaon 685 3 12 Before Kongla 635 5 Kongla 627 4 02 Kotni 610 4 02 Below Kishnapur 48 6 Ustem 44 6

River basin/Streams Seg. name Elevation (m) Stream order No of seg.

Ustem 2 44 6 Sonal 41 6 Cudcem 31 6 Velgeum 27 6 Khadaki 18 6 Khotode 14 6 Waghurme 14 7 Jamgaon 673 4 Bhandura Nerse 650 4 Panshet nala Talewadi 760 3 01 Bail nadi Bail nadi 721 3 01 Kalasa Satrem 112 3 06 Derodem 84 4 Nanodem 62 4 Kankumbi 750 3 Delta hotel 694 3 Checkpost 733 3 Kotryachi nadi Thane 85 5 04 Hedode 79 5 Naneli 65 5 Velus 49 5 Patwal Patwal 12 4 01 Ragada Vasant bandhara 135 1 06 Jambolim 21 4 Satpali 61 4 Panas 32 4 Shivade 16 4 Murmune 13 5 Dudhsagar Dudhsagar fall 146 4 06 Devachi kon 97 4 Dudhsagar juntion. 85 4 Cullem 74 5 Shigaon 58 5 Dabal 22 5 Caranzhol Cumtol 1 104 3 02 Cumtol 2 50 4 Karanjhol Karanjhol 83 4 01 ______

Table 2. 1b details of catchment area, direction of flow and the state in which the study is conducted (KN – Karnataka) River Headwater Downstream Catchment Direction State Sub-basin elevation (m) elevation (m) area (km2) of flow Mhadei 610-760 10 - 150 1332.02 West KN, Goa Malaprabha 640-720 - 0744.32 East KN Bhadra 750-930 - 225.12 East KN Tunga 630-880 - 160.00 East KN

Table 2. 1c sampling duration in each of the river sub-basin. ______Basin Duration of sampling Season Method Mesh size (cm) Bhadra May 2013 Dry Castnet 0.5 x 0.5, 1.1 x 1.1 June 2012, Nov- Dec 2013 Wet Castnet 0.5 x 0.5, 1.1 x 1.1 Tunga May 2013 Dry Castnet 0.5 x 0.5, 1.1 x 1.1 June 2012, June 2013, Dec 2013 Wet Castnet 0.5 x 0.5, 1.1 x 1.1 Malap. Apr 2011, Mar 2012, May 2014 Dry Castnet 0.5 x 0.5, 1.1 x 1.1 June 2011, Jan 2014 Wet Castnet 0.5 x 0.5, 1.1 x 1.1 Mhadei March 2012, April 2012, April 2013, Dry Castnet 0.5 x 0.5, 1.1 x 1.1 May 2014 June 2011, Jan 2014, Feb 2014 Wet Castnet 0.5 x 0.5, 1.1 x 1.1

Fish sampling Fish sampling was conducted with the help of cast net (0.5 cm x 0.5 cm and 1.1 cm x 1.1 cm mesh size). I relied on using a cast net over other fishing methods (gillnet and electrofishing etc) for the following reasons. First, it is a non-destructive fishing technique with the least fish mortality and is widely used in the tropical countries. It has been shown to capture most of the fish species present in a river habitat (Abraham and Kelkar 2012; Bhat, 2002) and can be operated by a single person with high sampling efficiency. It is also easily portable and enables more sampling points and locations. Most of the sampling segments were located in the deeper and undulating parts of the protected area which were only accessible by foot. A segment was sampled thoroughly with 18 casts for approximately 120 minutes to ensure that the catch represents most of the functional fish guilds including surface dwellers, mid- column dweller and bottom dwellers. I calculated fish sampling effort within stream segments (which included one or more mesohabitat units i.e. pool, run, riffle). These mesohabitat units were visually distinguished as areas for fish sampling as run consist of areas of stream stretches with gentle or fast water velocity with shallow depth, riffles as fast flowing water with turbulence due to submerged gravels and pebbles, pools are the areas with little or no water velocity and cascades are short waterfalls (Bhat, 2003; Hauer and Lamberti, 2007). Therefore the data collected in each mesohabitat unit was run (n=74), riffle (n=26), pool (n=51). Individuals of each fish were identified to the species level, measured (total length in cm) and released at the collected sites. Only unidentified individuals (1-2 per morpho-species) were collected, preserved in 90 % ethanol and subsequently transferred into 70 % ethanol in the laboratory for further identification. Standard textbooks following taxonomic key were used to identify species (Day, 1875; Daniels, 2002; Jayaram, 2010). Fish

24 taxonomists and fish biologists were also consulted for species identification, especially for some of the complex groups of fishes.

Habitat variables Data on habitat characteristics at each sampling segment were collected following the fish sampling. These included measurement of river width (m), river depth (m) three readings along the stream width, water temperature (0C), electrical conductivity (µS/cm), total dissolved solids (ppt), pH, canopy cover (%), and substratum (%). Portable water tracers were used to record electrical conductivity (Hanna Instruments HI98302), total dissolved solids (HI98304) and pH (HI96107). A water sample was also collected from the mid-section of each segment for the estimation of other water quality parameters in the laboratory including total hardness (mg/l), calcium hardness

(mg/l), free CO2 (mg/l), chlorides (mg/l), total alkalinity (mg/l), inorganic nitrates (mg/l) and inorganic phosphates (mg/l). All these samples were analyzed in the laboratory within 24-48 hrs. Water analysis was performed by following standard protocols (APHA, 2005; Trivedi and Goel, 1986). River bed substratum was visually categorized into proportions of rocks, boulders, sand, gravel and mud, leaves and wood. Canopy cover (%) was also visually estimated at each segment. Presence of riparian vegetation (dominant tree and shrubs and weeds) was noted. Previously sampled segments were re-sampled during a different season from the reverse direction to remove artefacts of direction of sampling.

Statistical Analysis In total, I sampled fish from 152 segments in 33 streams in four sub-basins to determine patterns of fish community organization. I calculated fish species diversity (richness and abundance) in each of the sampled segments. Species richness, the number of species recorded in each segment, was calculated based on species accumulation curves across four sub-basins and the number of individuals. Overall rank abundance was also calculated to examine the dominant species in the study area. All calculations were carried out in R statistic (R Development Core Team, 2011) using the package BiodiversityR (version 2.5-1). Generalized Linear Models (GLMs) were used to evaluate the relative influence of environmental drivers such as stream characteristics and water quality variables on fish richness and fish abundance.

I checked multicollinearity in predictor variables by calculating the variance inflation factor (VIF). VIF quantifies inflation in parameter estimates due to multicollinearity (Guisan et al., 2002). Variables that contributed to high variance were removed and those with VIF less than three were retained (Ott and Longneck, 2010). Model selection procedure based on AIC criteria was then used as a measure of information loss of each candidate model, with the best fitting model having the lowest AIC value (Crawley, 2007). I noted disturbance level in each of the sampled segments to test the effect of disturbance on species richness and abundance. Broadly, I defined a disturbance as any event that directly/indirectly affects fish diversity (richness and abundance). A level of ‘high disturbance’ was assigned to segments that exhibited either fishing pressure, sand-boulder mining, dynamiting, presence of pollution, presence of check dam or any hydrological barrier, medium disturbance to segment with water extraction for plantations, combined activities washing, bathing & vehicle washing, water diversion for hotel use, etc and ‘low disturbance’ to segments that lacked any significant level of disturbance but includes washing & bathing, livestock washing, vehicle washing, water extraction for drinking purpose (Appendix -1). Based on the intensity of disturbance and its likely impact on fish fauna, I ranked disturbance on a scale of 1 to 12, with low (1-4), medium (5-8) and high (9-12) disturbance rank.

Results Patterns in fish diversity (between river sub-basin and seasons) Fish species diversity assessed in four sub-basins revealed a total of 93 species belonging to nine orders and 18 families with 18,322 individuals. The order Cypriniformes was the richest (93% of the total orders) with 27 generas and 60 species followed by the orders Siluriformes and Perciformes (each with 6 genera and 13 species family Cyprinidae dominates among all families with 49 species followed by Balitoridae with 10 species and Cobitidae (1 species), (Appendix – Ib).

The cumulative species richness for the segments varies in each river sub-basin (Fig. 2. 2a). When the species richness of segments was compared, Malaprabha basin showed the highest observed species richness. For this basin, the species accumulation curve (SAC) did not reach an asymptote when the data for 15 segments

26 were added, suggesting that Malaprabha is a hyper diverse basin. Mhadei was the second most species rich basin, with the SAC stabilizing around forty species. The Tunga sub-basin also showed a similar trend, reaching an asymptote around 40 species. The Bhadra sub-basin, reached an asymptote quite early, and the species richness was estimated to be around 20. Malaprabha river sub-basin showed highest species richness (53 species in 15 sites) followed by Mhadei (47 species in 32 sites), Tunga (45 species in 52 sites) and Bhadra (24 Species in 41 sites), (Fig. 2. 2a). The cumulative species richness sampled across total number of individuals showed an interesting pattern. The maximum number of individuals were caught in Mhadei (7335) sub-basin alone, followed by Tunga (4447), Malaprabha sub-basin (4050) and Bhadra (2489) (Fig. 2b). The lower detection probabilities in capturing fish species richness and individual encounter rate using cast net has likely resulted in this variation in species richness and abundance.

My analyses also shows that wet season richness reached its asymptote earlier than the dry species richness (Fig. 2. 2c and d). The individual accumulation curves are also shown in Fig. 2. 2b and for the wet and dry season in Fig. 2. 2d. Species abundance in both dry and wet seasons accumulated steadily when more than 10,000 individuals were caught in the four sub-basins (Fig. 2. 2d).

Among all fishes, the five most abundant species belong to the family Cyprinidae, and include Devario malabaricus, Salmophasia boopis, Garra mullya, Rasbora daniconius and Barilius bakeri (Fig. 3a, b and Appendix -III). This family is by far the most abundant in the rivers of this region. Similar results were found for each of the river basins (Appendix III). Mhadei basin showed the highest fish abundance followed by Malaprabha, Bhadra and Tunga.

Figure 2. 2 (a). The species accumulation curve for each river sub-basin (b) species accumulation curve across number of individuals, (c) The species accumulation curve for dry and wet season using number of segments sampled and (d) pooled individuals from all the sampled segments.

Figure 2. 3 (a) dominant fish families (b) species abundance rank distribution showing five dominant fish species across four river sub-basins.

I also observed a hump-shaped relationship between the mean richness and mean abundance as a function of stream order (Table 2.1 d and 2.1 e; Fig. 2. 4a and b).

Table 2.1 d. Fish richness with mean, SD and SE across stream order ______Stream No of Fish Mean fish SE (SE) Order segments richness richness 1 2 2 0.143 0.34 (0.18) 2 3 3 0.215 0.42 (0.14) 3 39 10 0.715 0.46 (0.01) 4 49 9 0.643 0.49 (0.01) 5 11 11 0.786 0.42 (0.03) 6 8 11 0.786 0.42 (0.05) 7 1 2 0.143 0.36 (0.36)

Table 2.1 e. Fish abundance with mean, SD and SE across stream order ______Stream No of Fish Mean fish SE (SE) Order segments abundance abundance 1 2 23 0.248 1.04 (0.52) 2 3 141 1.516 8.15 (2.71) 3 39 4268 45.893 156.12 (4.004) 4 49 7901 84.957 276.32 (5.63) 5 11 4112 44.216 135.87 (12.35) 6 8 1827 19.646 60.93 (7.61) 7 1 50 0.538 2.65 (2.65)

Figure 2. 4. (a). The relationship between average fish species richness and stream orders (b) average species abundance and stream order.

The frequency distribution suggests that most species were rare while very a few species were abundant. Endemic species richness was higher in the middle stream orders especially in Tunga Mhadei and Bhadra compared to Malaprabha. Relatively sample size was less Malaprabha than other river sub-basins (Fig. 2.5).

Figure 2. 5. Endemic species richness in different stream orders (a) Bhadra (b) Tunga (c) Malaprabha (d) Mhadei.

Fish diversity (richness-abundance) and habitat variables The best model from GLMs showed that species richness was positively correlated with water temperature, free CO2, rocks and, gravels and negatively correlated with disturbance level, depth width ratio, boulders, cobbles, pebbles, sand, calcium hardness, inorganic nitrates and dissolved oxygen (Pseudo R2 = 0.55). Similarly, fish

30 abundance was positively correlated with dissolved oxygen, depth-width ratio, rocks, gravels, water temperature and free CO2 and negatively correlated with disturbance level, inorganic nitrates, calcium hardness, boulders, pebbles, and sand, respectively (Pseudo R2 = 0.59), (Table 2.2).

Table 2.2. Parameter estimates for fish species richness and fish species abundance.

______Response Coefficient Estimate (SE) CI (2.5%) CI (97.5%) Pseudo R2 AICc Fish richness (Intercept) *** 3.76 (0.68) 2.42 5.07 0.55 378.68 Disturbance Low*** -0.71 (0.13) -0.97 -0.45 Disturbance Med*** -0.33 (0.10) -0.52 -0.13 Free CO2** 0.006 (0.002) -0.001 0.010 Calcium hardness -0.001 (0.0008) -0.002 0.0005 Inorganic nitrates -0.23 (0.19) -0.62 0.14 Dissolved oxygen* -0.04 (0.02) 0.004 0.09 Depth width ratio -2.68 (1.66) -5.97 0.54 Water temperature* -0.05 (0.02) -0.11 -0.001 Electric conductivity*** 6.74 (1.98) 2.83 10.60 Rocks -0.0005 (0.003) -0.007 0.006 Boulders. -0.005 (0.003) -0.013 0.0009 Cobbles -0.04 (0.004) -0.013 0.005 Pebbles 0.001 (0.004) -0.006 0.010 Gravels 0.0004 (0.003) -0.005 0.006 Sand** -0.01 (0.004) -0.02 -0.003 Fish abundance Intercept*** 4.34 (0.17) 4.00 4.67 0.59 3234.5 Stream order*** -0.08 (0.01) -0.11 -0.06 Disturbance Low*** -1.61 (0.03) -1.68 -1.54 Disturbance Med*** -0.77 (0.02) -0.82 -0.72 Free CO2*** 0.01 (0.0007) 0.008 0.01 Chloride*** -0.004 (0.0004) -0.005 -0.003 Calcium hardness*** -0.001 (0.0002) -0.001 -0.001 Inorganic nitrates*** -0.19 (0.05) -0.300 -0.086 Dissolved oxygen*** 0.02 (0.005) 0.01 0.03 Depth width ratio*** 5.26 (0.42) 4.42 6.10 Water temperature*** 0.04 (0.006) 0.03 0.059 Rocks*** 0.01 (0.0009) 0.011 0.015

Boulders*** -0.01 (0.001) -0.01 -0.01 Pebbles*** -0.008 (0.001) -0.01 -0.006 Gravels*** 0.01 (0.001) 0.01 0.014 Sand*** -0.01 (0.001) -0.01 -0.009 Statistical significance (alpha = 0.05), p values * p <0.05, ** p <0.01, *** p <0.001).

Fish species richness was highest in the pool habitat followed by riffle and run (Fig. 2.6a). Fish abundance was highest in pool habitat followed by run and riffle (Fig. 2.6 b). One-way Analysis of Variance (ANOVA) was used to test the difference in variation in fish richness and fish abundance across major habitat types. Tukey’s test was used for post hoc comparison between different habitat types. However, there was no statistical difference in any of the habitat types.

Figure 2.6. Fish species richness (a) and abundance (b) across three habitat types.

Fish diversity (richness-abundance) and disturbance relationship About 13 of the segments in my data were classified as highly (‘more’) disturbed, and 139 were classified as least (‘less’) disturbed. My results demonstrate that streams with more disturbance were associated with higher species richness and the less disturbed segments showed greater fish abundance (Fig. 2.7a and b). ANOVA suggest that there was significant difference in the disturbance level both for richness (p =0.003) and abundance (p=0.07).

Figure 2.7 (a) Species richness and abundance (b) across disturbance levels.

Discussion Taxonomic richness (order, family, genera and species) This study provides a systematic examination of freshwater fish communities at different spatial scales (segment, habitat, stream order and sub-basins) within and across river sub-basins in the WG, thus enabling a comprehensive understanding of fish community organization. My results demonstrate that at large spatial scales, river size (sub-basin as well as stream orders) governed species richness (Fig. 2.2) and at local scales (segment level), stream characteristics and water quality variables predicted species richness and abundance (Table 2.2). Generally, water quality in the regulated sub-basins (Malaprabha and Mhadei) was degraded due to a variety of anthropogenic factors including land use patterns. In spite of this, these regulated sub- basins showed higher species richness than unregulated sub-basins (Bhadra and Tunga) where the “water quality” was better. The high richness observed in Malaprabha sub-basin could be primarily due to the presence of lentic habitat (pools) which are relatively more stable (due to the impoundment effect) than the run and

35 riffle habitats. The other possible reason for highest species richness in Malaprabha could be connectivity of other upstream habitats in the river with the back-waters of the Saundatti dam which has mitigated to some extent the decline in the discharge of the river. The next highest richness in the Mhadei and Tunga sub-basin could be attributed to the dense stream network yielding diverse river habitats and abundant food resource in the form of macro-invertebrate diversity (Subramanian et al., 2005; Ibrahampurkar, 2012). The higher richness of endemics found in the Tunga and Malaprabha basins could be due to the geomorphology and dense stream network that generates a diverse set of stream habitats that are also less disturbed from anthropogenic disturbances (Fig. 2.5).

Drivers of species diversity Results of the generalized linear models suggest that species richness was predicted by stream related variables and water quality parameters, a pattern commonly documented in stream fish ecology (Gido and Jackson, 2010; Matthews, 1998). Higher levels of calcium hardness and reduction in canopy cover negatively influenced fish richness while free CO2 and water temperature influenced richness positively. A higher amount of free CO2 (> 50 mg/l) was present in stagnated pools especially in Malaprabha and Mhadei. Bhadra and Tunga had relatively more runs and riffles which were higher in dissolved oxygen. Water temperature has a significant effect on aquatic life through its effect on metabolism (Chapman, 1996; Dallas, 2008). In this study, increase in water temperature was perhaps due to 1) the effect of dam (Lessard and Hayes, 2003) as higher water abstraction would have reduced flow downstream leading to shallower and disconnected pools, and 2) lack of adequate riparian vegetation cover especially along the regulated sections of rivers (Malaprabha and Mhadei).

The water chemistry in terms of its hardness and alkalinity has changed and these variables show increased levels at a few segments due to washing, bathing, and vehicle washing (Table 2) which might affect osmoregulation in fish making them become vulnerable to infectious diseases and thus increasing mortality (Trivedi and Goel, 1986; Chapman, 1996). The other reason for rise in calcium hardness level in regulated river sub-basin, could be the combined effect of water abstraction and

36 hydrological barriers by reducing the ratio of surface to ground water immediately downstream of flow (Wurts and Robert, 1992; Merz, 2013; Atkore et al., 2017). As a result, water column dependent positional guilds seem to have been affected (Atkore et al., in Prep). However, the diverse responses of diverse fish guild is difficult to assess to pinpoint the tolerance level of key fish groups. Future studies should focus on specific responses of these groups to different environmental factors.

Analyses of the temporal patterns also suggested that fish communities showed seasonality in their richness and abundance. Higher fish richness was found in the dry season especially in the pool habitats, which were a dominant feature of large rivers such as Malaprabha, Mhadei and Tunga (Fig. 2.6). The water abstraction at fewer sampled sites in Malaprabha river could have reduced the water depth forcing most fishes to congregate at such habitats during dry season. The findings of this study correspond with studies elsewhere where authors attributed the higher dry seasonal richness to the habitat loss, higher evaporation but not due to change in the water quality in selected river segments (Arthington et al., 2005). In my study, more disturbed segments had higher species richness. This may be because segments with more disturbance were dominated by pool habitats rather than run and riffle habitats. However, fish abundance was higher in less disturbed segments (with runs/riffles) suggesting these sites were occupied by fry and juveniles, especially during wet season.

This was partly confirmed in an exploratory analysis which indicated that the body size of dominant fishes in the regulated sub-basins was relatively larger (>12 cm) than non-regulated segments (Appendix III, Fig. 2. 4 and 2.5). These fishes and other bigger barbs (Hypselobarbus sp, Tor sp) in the regulated basins were exposed to higher fishing pressure (dynamite use and diverting or stopping water and agricultural/ town sewage) compared to the non-regulated river basins. As a result, many fishes (15-20 individuals) including a few endemic species (Osteochilius nashii and Labeo fimbriatus) showed signs of disease and deformities. Previous studies focusing on disturbance have shown mixed results with respect to richness and abundance in disturbed tropical streams (Kondoh, 2001; Dornelas, 2010; Deacon et al., 2015). When considering the impact of disturbance on fish communities, researchers found that natural and unimpacted rivers had more structured or non-

37 random species distribution than rivers with disturbance (Bhat and Magurran, 2007). A study conducted in the southern WG has shown that, higher degree of fish endemism was restricted to the mid-elevational zones that occur within protected area boundaries (Abraham and Kelkar, 2012). Such studies are becoming more important and question the efficacy of the current terrestrial protected area approach for conserving riverine biodiversity. The present study also corroborates these findings demonstrating that endemic fish richness was restricted to the middle stream orders which are mainly found in the non-regulated river systems. Habitat complexity was diverse in the Bhadra and Tunga sub-basins due to the presence of pool, run, riffle and cascades which provided suitable breeding ground for most of the macroinvertebrates (Subramanian et al., 2005). It is important to maintain adequate connectivity between the main and different stream orders (especially outside the protected area boundaries) that allow migratory fishes to make use of shallow, diverse habitats that are well oxygenated for spawning and feeding. The location of many hydrologic barriers in the middle stream orders is thus a major threat to the fish diversity of the WG. Thus, this study fills an important knowledge gap on the fish communities of the relatively less studied but diverse river systems of the central WG region.

Basin biogeography Among the studied river basins, three rivers were east flowing (Malaprabha, Bhadra and Tunga) and one river was west flowing (Mhadei) and these differences in flow may also account for differences in species richness as the catchments on either side of the WG differ in key biophysical characteristics. West flowing rivers are geologically young, have short courses with high gradient less sediment load, high water velocity, less catchment area (species-area effect), and lack an extensive network of tributaries compared to the east flowing rivers (Gunnell and Radhakrishnan, 2001). The steeper elevation gradient gives rise to numerous falls, cascades, and rapids which harbors rich endemic fish fauna (Mani, 1974; Johnson and Arunachalam, 2009) in west flowing rivers. The other factors that would have resulted in the variation in species richness are catchment size, channel morphology, land use history and basin biogeography. The channel morphology, especially substrate composition, in east and west flowing rivers in the headwater region was similar i.e. largely dominated by rocks, boulders, pebbles, gravels (making a suitable niche for

38 habitat specialist and endemic fishes) (Mani, 1974) but this composition reduced significantly at downstream sections which was dominated largely by sand and silt with more general and favoring widely distributed fish species.

Biogeographically, a few species were exclusively endemic to one or two river basins. For instance, the genera Barilius, Barbodes were restricted only to the Bhadra and Tunga river basins but were not found in the Malaprabha and Mhadei river basins. Similarly, certain species such as Glossogobius sp and Mugil sp were estuarine forms restricted only to the Mhadei river basin and not found in other river basins. Species such as Sicyopterus was also restricted to the upper reaches of Mhadei not found in the other study segments. Many large body-sized endemic species (Hypselobarbus dobsoni, H. thomassi, etc) were found only in the relatively disturbed segments during the monsoon season, in the Malaprabha river. Some of the relatively disturbed river segments (Katgali, Asoga and Rumewadi) in Malaprabha harbor potentially new species which indicates such areas need more conservation attention in terms of detailed species inventories. Many migratory species (eurytopic) fishes were captured in this river basin alone. Many of them are tolerant to wide disturbance regime.

Water sharing from one basin to other has been a contentious issue, especially when the rivers run between two or more states. For instance, Karnataka and Goa states are involved in a water sharing disputes over the waters of Mhadei and Malaprabha. Additionally, within the state of Karnataka, in order to meet drinking water needs of multiple cities, there is a plan to divert the headwater from Mhadei upstream tributaries (which is within the state) to the main Malaprabha river via an irrigation canal. This has elicited wide protests and conflicts already (http://indiatoday.intoday.in/story/river-dispute-karnataka-and-goa-headed-for-a- major-clash/1/461596.html). Further, many headstream sites in the basin have been demarcated for constructing new dams or barrages in the near future. Abstraction by irrigation canals in the past have already destroyed the habitats of many specialist fish species. Once this new canal becomes functional, it might homogenize habitats and associated fish assemblages across two basins, affecting endangering numerous habitat specialist, and endemic fish species.

Conservation status and endemism In terms of IUCN threat status (Molur et al., 2011), 16 species have not been evaluated in the conservation status report (Molur et al., 2011) simply due to lack of information from this region (see Appendix- I). In addition, of the species reported in the study, 8 species were classified as Endangered, 3 were Vulnerable and 2 were Near Threatened, 2 were critically endangered, 4 were Data Deficient and 58 were classified as Least Concern. This study describe a new fish species of Pethia striata from the study region (Atkore et al., 2015) and also reports a healthy population of the Critically Endangered fish species i.e. Barbodes wynaadensis and the presence of Hypselobarbus thomassi, a new distributional record for Labeo dussumieri filling an important knowledge gap in Indian fish taxonomy. Of the 93 species, about 52 species are endemic to the WG region and about 40 are non-endemic, but native to the Indian subcontinent. Mhadei and Malaprabha showed a greater degree of disturbance that has led to the destruction of natural river substratum, making segments deeper and dominated by sand, and agricultural sewage. In the case of Bhadra and Tunga, such instances of higher levels of disturbances are rare but they show a higher proportion of non-forest land cover, especially, plantations and agriculture. Additionally, water quality at certain segments was degraded due to washing activities, which might affect the fish population indirectly. In terms of river management, priority should be given to maintaining adequate connectivity between the main river channel and its distributaries. The types of disturbance that directly affect fish population should be regulated and effective ways of engaging local communities for river conservation should be promoted. Monitoring as well as conservation efforts should focus more on assessing endemic species diversity and community composition.

To conclude, this chapter demonstrates the importance of the four river basins (Mhadei, Malaprabha, Tunga and Bhadra) for fish diversity in the WGs. Malaprabha and Mhadei showed higher species richness, despite being regulated streams. However, abundance was higher in these non-regulated rivers including the Tunga basin. My results suggest that fish diversity including that of endemics is higher in large rivers with intact middle stream orders and the overall community structure is shaped by stream characteristics and water quality variables.

References

Abraham, Robin Kurian, and Nachiket Kelkar. 2012. Do terrestrial protected areas conserve freshwater fish diversity? Results from the Western Ghats of India. Oryx no. 46 (04):544-553. doi: doi:10.1017/S0030605311000937.

APHA. 2005. Standard Methods for Examination of Water and Wastewater. American Public Health Association WWA,. Washington, D.C. Pp. 1496.

Alabaster, J. S. and. L., R. S. 2013. Water quality criteria for freshwater fish (Second ed.): Elsevier. Pp. 261.

Araújo, Francisco Gerson, Benjamin Carvalho Teixeira Pinto, and Tatiana Pires Teixeira. 2008. Longitudinal patterns of fish assemblages in a large tropical river in southeastern Brazil: evaluating environmental influences and some concepts in river ecology. Hydrobiologia No. 618 (1):89. doi: 10.1007/s10750-008-9551-5.

Arthington, Angela H., Stephen R. Balcombe, Glenn A. Wilson, Martin C. Thoms, and Jon Marshall. 2005. Spatial and temporal variation in fish-assemblage structure in isolated waterholes during the 2001 dry season of an arid-zone floodplain river, Cooper Creek, Australia. Marine and Freshwater Research No. 56 (1):25-35. doi: http://dx.doi.org/10.1071/MF04111.

Astorga, A., Death, R., Death, F. and Paavola, R. 2014. Habitat heterogeneity drives the geographical distribution of beta diversity: the case of New Zealand stream invertebrates. Ecology and evolution doi: 10.1002/ece3.1124.

Atkore, V. M., J. D. Marcus Knight, K. Rema Devi, and J. Krishnaswamy. 2015. A New Species of Pethia from the Western Ghats, India (Teleostei: Cyprinidae). Copeia No. 103 (2):290-296. doi: 10.1643/OT-12-172.

Atkore, V., N. Kelkar and J. Krishnaswamy 2017. Assesing the recovery of fish assemblages downstream of hydrological barriers in India's Western Ghats.River Research and Application. doi.10.1002/rra.3163.

Atkore V., N. Kelkar., S. Badiger., K. Shanker and J. Krishnaswamy. 2017. River barriers and environmental variables influence fish guild responses in the Western Ghats of India. In Prep.

Bhat, A. 2002. A study of the diversity and ecology of freshwater ﬁshes of four river systems of the Uttara Kannada District, Karnataka, India. A PhD thesis submitted to Centre for Ecological Sciences, Indian Institute of Science, Bangalore. Pp. 178.

Bhat, A. 2003. Diversity and Composition of Freshwater Fishes in River Systems of Central Western Ghats, India. Environmental Biology of Fishes No. 68 (1):25- 38. doi: 10.1023/a:1026017119070.

Bhat, A. 2004. Patterns in the distribution of freshwater fishes in rivers of Central Western Ghats, India and their associations with environmental gradients. Hydrobiologia No. 529:83-97. doi: 1007/s10750-004-4949-1. Bhat, A., and A. E. Magurran. 2007. Does disturbance affect the structure of tropical fish assemblages? A test using null models. Journal of Fish Biology No. 70 (2):623-629. doi: 10.1111/j.1095-8649.2006.01252.x.

Bhatt, J. P., K. Manish and M. K. Pandit. 2012. Elevation gradients in fish diversity in the Himalaya: Water discharge is the key driver of ditribution patterns. PLoS ONE, 79(9). doi: 10.1371/journal.pone.0046237

Brasher, A. M. D. 2003. Impacts of Human Disturbances on Biotic Communities in Hawaiian Streams. BioScience, 53(11), 1052-1060. doi: 10.1641/0006- 3568(2003)053[1052:iohdob]2.0.co;2

Bunn, E. Stuart, and H. Angela Arthington. 2002. Basic Principles and Ecological Consequences of Altered Flow Regimes for Aquatic Biodiversity. Environmental Management No. 30 (4):492-507. doi: 10.1007/s00267-002- 2737-0.

Burridge, Christopher P., Dave Craw, Daniel C. Jack, Tania M. King, and Jonathan M. Waters. 2008. Does fish ecology predict dispersal across a river drainage divide? Evolution No. 62 (6):1484-1499. doi: 10.1111/j.1558- 5646.2008.00377.x.

Chapman, D. 1996. Water quality assessment - a guide to the use of biota, sediments and water in environmental monitoring (D. Chapman Ed. Vol. Second): Chapman & Hall on behalf of UNESCO, WHO and UNEP.Pp 653.

Chea, R., Grenouillet, G., and Lek, S. 2016a. Evidence of Water Quality Degradation in Lower Mekong Basin Revealed by Self-Organizing Map. PLoS ONE, 11(1), e0145527. doi: 10.1371/journal.pone.0145527

Chea, R., Lek, S., Ngor, P., and Grenouillet, G. 2016b. Large-scale patterns of fish diversity and assemblage structure in the longest tropical river in Asia. Ecology of Freshwater Fish, n/a-n/a. doi: 10.1111/eff.12301

Chakrabarty, M., and Homechaudhuri, S. 2013. Fish guild structure along a longitudinally-determined ecological zonation of Teesta, an eastern Himalayan river in West Bengal, India. Arxius de Miscel-lania Zoologica, 11, 196-213.

Crawley, M. J. 2007. The R Book. The Atrium Southern Gate, Chichester, West Sussex PO 19 8SQ, England : John & Wiley Sons, Ltd. Pp. 877.

Day, F. 1875. The Fishes of India: Beinga natural history of fishes known to inhibit the sea and freshwaters of India, Burma and Ceylon. Vol. I & II. London :B. Quaritch,1876 [i.e. 1875]-1878.

Dallas, Helen. 2008. Water temperature and riverine ecosystems: An overview of knowledge and approaches for assessing biotic responses with special reference to South Africa. Water SA No. 34 (3).

Daniels, R.J.R. 2002. Freshwater fishes of Peninsular India. Hyderabad: University Press India (Pvt) Ltd.Pp. 320.

Deacon, Amy E., Hideyasu Shimadzu, Maria Dornelas, Indar W. Ramnarine, and Anne E. Magurran. 2015. From species to communities: the signature of recreational use on a tropical river ecosystem. Ecology and Evolution No. 5 (23):5561-5572. doi: 10.1002/ece3.1800.

DeFries, Ruth, Francesco Rovero, Patricia Wright, Jorge Ahumada, Sandy Andelman, Katrina Brandon, Jan Dempewolf, Andrew Hansen, Jenny Hewson, and Jianguo Liu. 2010. From plot to landscape scale: linking tropical biodiversity measurements across spatial scales. Frontiers in Ecology and the Environment No. 8 (3):153-160. doi: 10.1890/080104.

Dornelas, Maria. 2010. Disturbance and change in biodiversity. Philosophical Transactions of the Royal Society B: Biological Sciences No. 365 (1558):3719-3727. doi: 10.1098/rstb.2010.0295.

Edds, David R. 1993. Fish Assemblage Structure and Environmental Correlates in Nepal's Gandaki River. Copeia No. 1993 (1):48-60. doi: 10.2307/1446294.

Gaston, Kevin J. 2000. Global patterns in biodiversity. Nature No. 405 (6783):220- 227.

Gido, K., and Donald A. Jackson. 2010. Community Ecology of Stream Fishes: Concepts, Approaches, and Techniques. Bethseda, Maryland, USA: American Fisheries Society Sympposium 73. Pp.655.

Gorman, Owen T., and James R. Karr. 1978. Habitat Structure and Stream Fish Communities. Ecology No. 59 (3):507-515. doi: 10.2307/1936581.

Gunell, Y., and Radhakrishnan, B. P. 2001. Sahyadri: the great escarpment of the Indian subcontinent (patterns of landscape development in the Western Ghats) (eds.) Geological Society of India, Bangalore.Vol. I. Pp 1053.

Grenouillet, Gaël, Didier Pont, and Knut L. Seip. 2002. Abundance and species richness as a function of food resources and vegetation structure: juvenile fish assemblages in rivers. Ecography No. 25 (6):641-650. doi: 10.1034/j.1600- 0587.2002.250601.x.

Guisan, Antoine, Thomas C. Edwards Jr, and Trevor Hastie. 2002. Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecological Modelling No. 157 (2–3):89-100. doi: http://dx.doi.org/10.1016/S0304-3800(02)00204-1.

Hauer, F.R and G. Lamberti. 2007. Methods in stream ecology. Edited by F.R and G. Lamberti Hauer. Vol. 2nd: Academic Press.Pp.896.

Heller E, J.M. Rhemtulla, S. Lele, M. Kalacska, S. Badiger, R. Sengupta and, and N. Ramankutty. 2012. Mapping crop types, irrigated areas and cropping intensities in heterogeous landscapes of southern India using multi-temporal medium resoultion imagery: Implications for assessing water use in agriculture. Photogrammetric engineering and Remote Sensing No. 78 (8):815-827.

Holomuzki, Joseph R., Jack W. Feminella, and Mary E. Power. 2010. Biotic interactions in freshwater benthic habitats. Journal of the North American Benthological Society No. 29 (1):220-244. doi: 10.1899/08-044.1.

Hugueny, B., T. Oberdorff., and P. A. Tedescco. 2010. Community ecology of river fishes: a large-scale perspective. In 'Community Ecology of Stream Fishes: Concepts, Approaches and Techniques'. In American Fisheries Society Symposium, edited by and Donald A. Jackson Keith B. Gido: The American Fisheries Society.Pp 29-62 Ibrahampurkar, M. M. 2012. Hydrological and hydrogeological evaluation of Mhadei river watershed in Goa and Karnataka. A Ph.D thesis submitted to The Department of Earth Science. Goa University, Panjim. Pp 282. Jackson, Donald A., Pedro R. Peres-Neto, and Julian D. Olden. 2001. What controls who is where in freshwater fish communities Â– the roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries and Aquatic Sciences No. 58:157-170. Jardine Timothy D., Nicholas R. Bond, Michele A. Burford, Mark J. Kennard, Douglas P. Ward, Peter Bayliss, Peter M. Davies, Michael M. Douglas,Stephen K. Hamilton, John M. Melack, Robert J. Naiman, Neil E. Pettit, Bradley J. Pusey, Danielle M. Warfe, and Stuart E. Bunn. 2015. Does flood rhythm drive ecosystem responses in tropical riverscapes? Ecology No. 96 (3):684–692.

Jayaram, K. C. 2010. The Frehswater Fishes of Indian Region. Delhi: Narendra Publishing House.

Johnson J. A. and Arunachalam, M. 2009. Diversity, distribution and assemblage structure of fishes in streams of southern Western Ghats, India. Journal of Threatened Taxa No. 1 (10):507-513.

Johnson, J. A and Muthukumarasamy Arunachalam. 2010. Habitat use of fishes in streams of Kalakad Mundanthurai Tiger Reserve, India. International Journal of Ecology and Development No. 17 (F 10):34-47.

Kasturirangan, K. 2013. Report of the High Level Working Group on Western Ghats. Ministry of Environment and Forests, Government of India. Part I & II.

Khedkar, G. D., Lutzky, S., Rathod, S., Kalyankar, A., and David, L. 2014. A dual role of dams in fragmentation and support of fish diversity across the Godavari River basin in India. Ecohydrology, 7(6), 1560-1573. doi: 10.1002/eco.1470

Kondoh, Michio. 2001. Unifying the relationships of species richness to productivity and disturbance. Proceedings of the Royal Society of London. Series B: Biological Sciences No. 268 (1464):269-271. doi: 10.1098/rspb.2000.1384.

Lessard, J. L., and Hayes, D. B. 2003. Effects of elevated water temperature on fish and macroinvertebrate communities below small dams. River Research and Applications, 19(7), 721-732. doi: 10.1002/rra.713

Mani, M. S. 1974. Ecology and Biogeography in India. The Hague: W Junk Publishers. Pp.775.

Macnaughton, C. J., Senay, C., Dolinsek, I., Bourque, G., Maheu, A., Lanthier, G., . . . Boisclair, D. 2016. Using fish guilds to assess community responses to temperature and flow regimes in unregulated and regulated Canadian rivers. Freshwater Biology, 61(10), 1759-1772. doi: 10.1111/fwb.12815

Magurran, Anne E., and A. T. Phillip Dawn. 2001. Implications of Species Loss in Freshwater Fish Assemblages. Ecography No. 24 (6):645-650.

Malkhede, S. K. 2003. Working Plan for Belgaum Forest Division (2002-2003 to 2011-2012). Belgaum, Karnataka state, India: Belgaum Forest Division, Belgaum.

Maltchik L., L. E. K. Lanes., C Stenert and E. S. F. Medeiros. 2010. Species-area relationship and environmental predictors of fish communities in coastal freshwater wetlands of southern Brazil. Environmental Biology of Fishes No. 88 (25-35).

Matthews, M. J. 1998. Patterns in freshwater fish ecology. Massachusetts, USA.: Kluwer Academic publisher.Pp. 670.

Mattos, T. M., Costa, M. R., Pinto, B. C. T., Borges, J. L. C., and Araújo, F. G. 2014. To what extent are the fish compositions of a regulated river related to physico-chemical variables and habitat structure? Environmental Biology of Fishes, 97(6), 717-730. doi: 10.1007/s10641-013-0175-x

Menni, R., Gómez, S., and Armengol, F. 1996. Subtle relationships: freshwater fishes and water chemistry in southern South America. Hydrobiologia, 328(3), 173- 197. doi: 10.1007/BF00017629

Merz, S. K. 2013. Characterizing the relationship between water quality and water quantity (E. Department of Sustainability, Water, Population and Communities, Trans.) Pp. 139. St Leonards NSW 2065 Australia.

Molur, S., K. G. Smith, B. A. Daniel and W. R. T. Darwall. 2011. The Status and distribution of freshwater biodiversity in the Western Ghats, India. Cambridge,

UK and Gland, Switzerland: IUCN, and Coimbatore, India: Zoo Outreach Organisation.: IUCN.

Oberdorff T., Guegan, J-F., Hugueny, B. 1995. Global scale patterns of fish species richness in rivers. Ecography. Vol 18 No 4:345-352.

Oberdorff, Thierry, Pablo A. Tedesco, Bernard Hugueny, Fabien Leprieur, Olivier Beauchard, Sébastien Brosse, and Hans H. Dürr. 2011. Global and Regional Patterns in Riverine Fish Species Richness: A Review. International Journal of Ecology No. 2011:12. doi: 10.1155/2011/967631.

Olden, Julian D., Mark J. Kennard, Fabien Leprieur, Pablo A. Tedesco, Kirk O. Winemiller, and Emili García-Berthou. 2010. Conservation biogeography of freshwater fishes: recent progress and future challenges. Diversity and Distributions No. 16 (3):496-513. doi: 10.1111/j.1472-4642.2010.00655.x.

Ott, R. Lyman, and Michael T. Longneck. 2010. An introduction to statistical methods and data analysis. 6 ed. Texas A and M University.

Poiani, K. A., Richter, B. D., Anderson, M. G., and Richter, H. E. 2000. Biodiversity Conservation at Multiple Scales: Functional Sites, Landscapes, and Networks. BioScience, 50 (2), 133-146. doi: 10.1641/0006- 3568(2000)050[0133:BCAMSF] 2.3.CO;2

Pool, T. K., Olden, J. D., Whittier, J. B., and Paukert, C. P. 2010. Environmental drivers of fish functional diversity and composition in the Lower Colorado River Basin. Canadian Journal of Fisheries and Aquatic Sciences, 67(11), 1791-1807. doi: 10.1139/F10-095

Pusey, Bradley J, Angela H Arthington, and Martin G Read. 1998. Freshwater fishes of the Burdekin River, Australia: biogeography, history and spatial variation in community structure. Environmental Biology of Fishes No. 53 (3):303-318. doi: 10.1023/A:1007468417947.

Raghavan, Rajeev, G. Prasad, P. H. Anvar Ali, and Benno Pereira. 2008. Fish fauna of Chalakudy River, part of Western Ghats biodiversity hotspot, Kerala, India: patterns of distribution, threats and conservation needs. Biodiversity and Conservation No. 17 (13):3119-3131. doi: 10.1007/s10531-007-9293-0.

R Development Core Team. 2011. R: A Language and Environment for Statistical Computing. Vienna, Austria: the R Foundation for Statistical Computing. ISBN 3-900051-07-0. Available onlime at http:/www.R-project.org/.

Ricklefs, R. E., and Dolph, S. 1993. Species Diversity in Ecological Communities: University of Chicago Press. Pp. 414.

Rosenzweig, M. L. 1995. Species diversity in space and time. Cambridge.: Cambridge University Press. Pp.436.

Sakaris, P. C. 2013. A review of the effects of hydrologic alteration on fisheries and biodiversity and the management and conservation of natural resources in regulated river systems. in Paul M. Bradley (editor). Current perspectives in contaminant hydrology and water resource sustainability.

Schlosser, Isaac J. 1991. Stream Fish Ecology: A Landscape Perspective. BioScience No. 41 (10):704-712. doi: 10.2307/1311765. Schwartz, M. W. 1999. Choosing the Appropriate Scale of Reserves for Conservation. Annual Review of Ecology and Systematics, 30(1), 83-108. doi: 10.1146/annurev.ecolsys.30.1.83

Sheldon, F., and Fellows, C. S. 2010. Water quality in two Australian dryland rivers: spatial and temporal variability and the role of flow. Marine and Freshwater Research, 61(8), 864-874. doi: http://dx.doi.org/10.1071/MF09289

Subramanian, K. A., K. G. Sivaramakrishnan, and Madhav Gadgil. 2005. Impact of riparian land use on stream insects of Kudremukh National Park, Karnataka state, India. Journal of Insect Science No. 5:49.

Tedesco Pablo, A., Leprieur Fabien, Hugueny Bernard, Brosse Sébastien, H. Dürr Hans, Beauchard Olivier, Busson Frédéric, and Oberdorff Thierry. 2012. Patterns and processes of global riverine fish endemism. Global Ecology and Biogeography. doi: 10.1111/j.1466-8238.2011.00749.x.

Tondato, K. K. and Y. R. Suarez. 2010. Temporal changes in fish species composition of headwater streams of the upper Praguay and Parana basins, Brazil. Acta Scientiarum. Biologcal Sciences No. 32 (3). doi:10.4025/actascibiolsci.v32i3.6160.

Trivedi, R.K. and Goel P.K. 1986. Chemical and Biological methods for water pollution studies: Environmental Publications, India.

Unmack, P T. 2001. Biogeography of Australian Freshwater fishes.Journal of Biogeography No. 28 (9):1053-1089. van Jaarsveld, A. S., Freitag, S., Chown, S. L., Muller, C., Koch, S., Hull, H., . . . and Scholtz, C. H. 1998. Biodiversity Assessment and Conservation Strategies. Science, 279 (5359), 2106-2108. doi: 10.1126/science.279.5359.2106

Vokoun, J. C., and Y. Kanno. 2009. Evaluating effects of water withdrawls and impoundments on fish assemblages in Connecticut streams. (pp. 31). University of Connecticut, Storrs: Connecticut Department of Environmental Protection.

Wikramnayake, E. D. 1990. Ecomorphology and biogeography of a tropical stream fish assemblage: evolution of assemblage structure. Ecology No. 71 (5):1756- 1764.

Winemiller, K. O., and Jepsen, D. B. 1998. Effects of seasonality and fish movement on tropical river food webs. Journal of Fish Biology, 53, 267-296. doi: 10.1111/j.1095-8649.1998.tb01032.x

Wurts, W.A. , and M. Durborow Robert. 1992. Interactions of pH, carbon dioxide, alkalinity and hardness in fish ponds. Southern Regional Aquaculture Center, USA (464):1-4. Yap, S. Y. 2002. On the distributional patterns of Southeast-East Asian freshwater fish and their history. Journal of Biogeography No. 29 (9):1187-1199. doi: 10.1046/j.1365-2699.2002.00771.x.

Chapter 3 Determinants of native fish species turnover in the Western Ghats, India

Introduction

Biodiversity at the global scales is concentrated in tropical regions but determining drivers of diversity for specific groups within the tropics is an area of active investigation (Arthington et al., 2010; Vorosmarty et al., 2010; Boyero et al., 2011; Dudgeon, 2011; McGill et al., 2015). However, little is known about how biodiversity, especially, fish communities change along an environmental gradient. This is interesting both from a biogeographical perspective, as it offers insights into the mechanisms by which species are accumulated (nestedness, turnover, colonization, extinction etc) and from the perspective of understanding the effect of river regulation due to dams and other water diversion schemes. Community change across environmental gradients has been understood by assessing species turnover by using beta diversity measures (Magurran, 2004; Alfonsi et al., 2016).

Whittaker (1960) made the distinction between α- and β-diversity. He defined α- diversity as the property of a defined spatial unit, while β-diversity reflects biotic change or species replacement as from one spatial unit to the next. The concept of β- diversity or turnover can include changes in diversity over space and time. Examining species compositional change along any ecological gradient has become an important aspect of monitoring biological diversity (Magurran, 2004), with important applications such as designing and managing protected areas (Wiersma and Urban, 2005; Krishnaswamy et al., 2009). The species compositional change reflects both historical and environmental events (Philippi et al., 1998). Recent studies have conceptualized and refined the idea of species turnover (Whittaker, 1960; Wilson, 1984; Borcard et al., 1992; Lande, 1996; Gering et al., 2003; Koleff, 2003) with clear illustrations of methodology (Jost, 2007; McDonald and Hillebrand, 2007; Baselga, 2010; 2012, 2013; Soininen, 2010; Tuomisto, 2010, Veech, 2010; Anderson et al., 2011; Robert and Meelis, 2014). Despite this, there exists an ambiguity regarding the drivers of the species turnover across multiple spatial scales.

River systems are linear net-works that are nested at multiple spatial scales from head-waters to large river channels. Therefore, they offer unique opportunities to test spatial structure in aquatic communities (Al-Shami et al., 2013; Hauer and Lamberti, 2007). Beta-diversity is known to depend on spatial scale since geographic proximity may be correlated with species similarity (Boyero, 2003; 2011; Heino et al., 2015). Fish species diversity is influenced by biotic and abiotic processes that function across various scales of space and time in riverine systems (Pegg, 2007). Some of these processes operate at very coarse temporal and spatial scales. At larger scales, biotic processes such as speciation and dispersal associated with geomorphology and tectonics strongly influence the pattern of fish diversity within and among river systems. However, at finer spatial scales, the flow of water across the fluvial landscape influences fish species diversity, as it can have strong effects on biotic factors such as dispersal and trophic interactions (predation etc) (Pegg, 2007). Abiotic factors related to local geology, and river-channel morphology are largely determined by the interaction between the flow regimes of a river which play a crucial role in determining the types of suitable habitat available for riverine fishes throughout the various stages of their life history (Bunn and Arthington, 2002; Pegg, 2007).

River valley projects are known to transform natural landscapes and the riverscapes within them which probably affect river and associated freshwater ecosystems in many ways (Dudgeon, 2000 a, b; Vorosmarty et al., 2010). Their impacts are often reflected in changing the spatial patterns of fish assemblages (Dudgeon, 2011). Particularly, dam construction can fragment basins, affect fish assemblages directly by eliminating or reducing movement of fishes, which then leads to reduction of upstream/downstream species richness especially that of migratory species (Gillette, 2005). Dams also alter connectivity (Pringle, 2001) and have substantial cumulative effects on fish biodiversity at a regional scale. The effect of dams was shown to last for more than a decade after its construction, which led to changes in the fish guild composition at downstream sections of a river in Brazil (Vasconcelos et al., 2014a). On the other hand, homogenization in river systems due to the impact of invasive species and channel modification (Bertuzzo, 2009) will have an effect on fish species composition.

Studies on species composition and turnover are relatively fewer in Indian river systems. Previous studies on species turnover in Indian river systems have suggested

50 that species compositions across upstream/downstream reaches are more similar to each other than the upstream and downstream reaches in the same river (Bhat, 2003). Understanding of spatial turnover of stream fishes in the WG is of considerable interest to biogeographers and also serves as an evaluation of the adequacy of the PA network in the protection of fish diversity. Further, it can enhance our understanding of how fish species composition is affected in altered river systems.

In this chapter, I chose two adjacent river sub-basins from the Western Ghats (WG) to answer two questions: 1) What drives the fish species turnover between river segments within a sub-basins? 2) How does fish species turnover vary across adjacent river basins? I hypothesize that the similarity in environmental variables will determine fish species composition in addition to geographic distance (Leprieur et al., 2009).

Materials and Methods Study Site The study was conducted in the WG Biodiversity Hotspot and a UNESCO World Heritage site in India (Das et al., 2006; Kasturirangan, 2013). I chose Bhadra and Tunga rivers that are separated by a hill range situated at 1160 m near Gangamoola in the Kudremukh National Park and have similar topographic and climatic features. Both these rivers are tributaries of Krishna river: a major east-flowing river in peninsular India (Fig. 3.1). Bhadra and Tunga catchments are dominated by tropical wet evergreen forest and montane habitats which comprise stunted evergreen forests (sholas) in a grassland matrix. These sub-basins are fed by numerous perennial streams and exhibits a wide array of stream habitat types such as runs, pools, cascades and riffles.

Sampling strategy In Bhadra, I sampled stream segments across an elevation range from 600 m to 1160 m. The sampled segments covered a range of riverine habitats that are essential for fish fauna. In Tunga, the elevation range covered was between 600 m and 900 m. The major habitats sampled were pool, run, and riffle. I defined a river segment as my sampling unit, which was a stretch of 100-150 m length.

Figure 3. 1. Sampling locations in Tunga and Bhadra river sub-basins.

These segments were nested under different stream orders and spaced approximately 300-500 m distance apart. I used only wet season (June and November-December, 2013) across two sub-basins for this analysis considering the logistical constraints.

Fish sampling Fish sampling was conducted with the help of castnets (0.5 cm x 0.5 cm and 1.1 cm x 1.1 cm mesh size) as well as with modified fish census method (Arunachalam, 2000; Deacon et al., 2015) at certain river sections where cast netting was not feasible. Each river segment was sampled thoroughly for about 120 minutes, depending on feasibility. Fishes were identified to the species level in the field and were released live at the captured locations after measuring body length in cm. River habitat characteristics data was collected following fish sampling in each sampling segment. This included river width (m), river depth (m), water temperature (0C), electrical conductivity (µS/cm) (model HI98302), total dissolved solids (ppt) (model HI98304), pH (model HI96107), canopy cover (%), and substratum (%). Portable water tracers were used to record electrical conductivity (µS/cm), total dissolved solids (ppt) and pH (Hanna Instruments). River bed substratum was visually categorized into rocks, boulders, sands, gravels and mud, leaves and woods (%). Canopy cover (%) was also noted down at each segment. Segments were re-sampled over two wet seasons from the reverse direction to prevent sampling sequence artefacts. Water samples were also collected from the midsection of each segment for laboratory estimation of other water quality parameters such as total hardness (mg/l), calcium hardness (mg/l), free

CO2 (mg/l), chlorides (mg/l), total alkalinity (mg/l), inorganic nitrates (mg/l) and inorganic phosphates (mg/l). Analysis of water quality parameters was carried out following the standard (APHA; Trivedi and Goel, 1986).

Statistical Analysis All analyses were performed in R statistical software (R Development Core Team, 2012) using the ‘ecodist’ package (Goslee and Urban, 2007).

Species composition and turnover I calculated dissimilarity in species composition (i.e. species turnover) across segments using Bray-Curtis index (that uses relative abundance data). For stream characteristics (variables), I used Euclidean distance between sites to analyze variability in freshwater fish species and environmental covariates (Jackson et al., 2001a). Association between geographical proximity and fish faunal composition was calculated by using Mantel’s test. As one would expect, the greater the geographic

53 distance between regions, the less is biological similarity (Goslee and Urban, 2007). A monotonic decrease in Mantel’s correlogram would indicate effect of space and the periodic pattern would indicate the influence of habitat configuration rather than pure spatial proximity. I also calculated partial Mantel correlation for all the measured variables to check the influence of variables on dissimilarity in species composition.

Reference composition I estimated pair-wise dissimilarity between segments and also estimated dissimilarity with respect to a hypothetical reference composition to enable analyses that required a continuous measures of compositional turnover across spatial units. I calculated a hypothetically diverse reference composition by combining all the species encountered in two river sub-basin. I used incidence data both for total as well as for endemic species. I then took sum of all the species for all sampled segment across two basins. Therefore every segment in a study region is related to the reference composition. Dissimilarity was calculated between each segment and the newly created reference composition. Since all the segments were sampled from upstream to downstream directions spaced at 300-500 m distance, paired difference for every segment was calculated by taking an average between immediate upstream and downstream segments of every segment.

Spatial analysis The dissimilarity was mapped using Arc Map GIS software across different segments. Mapping helped in visually interpreting the degree of spatial dependence present in all the sampled segments. Association between geographical proximity and fish faunal composition was calculated using Mantel’s tests and correlogram.

Habitat variables I also tested for spatial autocorrelation using Mantel’s correlograms in stream ecological variables using Euclidean distance between sites for fish species and ecological covariates. To identify if any specific covariates varied in space and predicted species turnover, I used Mantel correlations between similarity and habitat covariates.

Results Pattern of species composition In total, I recorded 40 species in Bhadra and Tunga river basins with 5004 individuals. I recorded a total of 20 species with 1829 individuals (36.55%) in Bhadra and 39 species with 3175 individuals (63.44%) in Tunga (Appendix – 3.1). The Mantel’s correlogram indicated a sine-hole pattern, which represented high turnover at small spatial distance (up to 5 km) across Bhadra Tunga sub-basin (Fig.3. 2). Species turnover or dissimilarity in species composition was strongly correlated with the geographic distance (Mantel’s r = 0.16, p = 0.0001). Similarly, I calculated Mantel’s correlogram for individual river basin. In Bhadra, species turnover effect was weakly correlated or uncorrelated with geographic distance (r = -0.013, p = 0.97), it becomes positive after 5 km (Fig. 3. 3).

Figure 3.2. Relationship between species dissimilarity as a function of geographical distance for Bhadra Tunga combined. Correlogram shows a sine-hole pattern where sites which are closer are also geographically similar (indicated by filled circles) than sites that are farther apart. Species turnover is high up to a distance of 15 km and strongly correlated with geographical distance (Mantel’s r = 0.16, p=0.0001) and the effect of species turnover becomes positive after a distance of 10 km.

Figure 3.3. Correlogram showing the relationship between pairwise dissimilarity and geographical distance in Bhadra. Species turnover effect is weakly correlated with geographic distance (r =-0.013, p = 0.97) and it becomes positive after 5 km distance in Bhadra.

In the case of Tunga, species turnover was high at smaller spatial scales (up to a distance of 1-2 km). Species turnover was strongly associated with geographical distance (Mantel r = -0.21, p = 0.002), Fig. 3. 4).

Figure 3. 4. Correlogram showing the relationship between pairwise dissimilarity and geographical distance in Tunga. Compositional dissimilarity (species turnover) showed a sine-hole pattern with high species turnover at smaller spatial scales (up to a distance of 1-2 km). In other words, compositional dissimilarity was strongly associated with geographical distance (Mantel r=-0.21, p =0.002).

Drivers of species composition Species turnover in Bhadra (up to 5 km) (Fig. 3. 3) and (up to 6 km) in Tunga (Fig. 3. 4) indicated the greater habitat connectivity and fish mobility. The high turnover was attributed to local factors (variation in habitat quality and biophysical parameters) rather than effects of spatial proximity. Mantel r was significant and positively correlated with geographical distance (0.16), rocks (0.19), boulders (0.11), pebbles (0.13) and water temperature (0.11) and negatively correlated with electrical conductivity in combined Bhadra-Tunga sub-basins (Table 3. 1).

Table 3. 1. Mantel’s test shows the relationship between pairwise dissimilarity in species composition and dissimilarity in ecological covariates for Bhadra Tunga sub- basin (combined). ______Covariates Mantel r (p value) 2.5% llim 97.5% ulim Geographical distance*** 0.16 (0.0001) 0.14 0.19 Rocks** 0.19 (0.004) 0.14 0.22 Boulders** 0.11 (0.006) 0.07 0.15 Pebbles** 0.13 (0.003) 0.10 0.17 Stream order 0.09 (0.02) 0.05 0.12 Slope 0.06 (0.07) 0.037 0.100 Depth width ratio 0.081 (0.87) 0.04 0.14 Water temperature* 0.11 (0.03) 0.08 0.14 Electrical conductivity -0.06 (0.87) -0.11 -0.03 pH 0.08 (0.11) 0.04 0.10

Partial Mantel r 0.04 (0.14) 0.01 0.08

*Statistical significance alpha = 0.05: (p-values * p <0.05, ** p <0.01, *** p <0.001)

In Bhadra, Mantel’s r was significant and positively correlated with stream order (0.15), rocks (0.15), and canopy cover (0.14) while it was negatively correlated with depth- width, and EC indicating that the overall richness was low (20 species) but with different species composition (see Table 2 for more detailed species list). The distinct composition was mainly due to the endemic and habitat specialist fish such as, Garra stenorhynchus. Similarly, in Tunga, the Mantel r was significant and positively correlated with gravels rocks (0.31), boulders (0.21), stream order (0.20), and depth- width ratio (0.22), water temperature (0.17) and pH (0.16) etc (see Table 3. 3 for detail).

Table 3. 2. Mantel’s test shows the relationship between pairwise dissimilarity in species composition and dissimilarity in ecological covariates in Bhadra sub-basin. Species turnover was driven mainly by stream characteristics (stream order, canopy cover) and local habitat effects (rocks). ______Covariates Mantel r (p value) 2.5% llim 97.5% ulim Geographical distance -0.13 (0.97) -0.17 -0.09 Stream order** 0.15 (0.03) 0.08 0.22 Canopy cover** 0.14 (0.08) 0.10 0.18 Rocks** 0.15 (0.08) 0.10 0.20 Slope 0.09 (0.11) 0.02 0.16 Depth width ratio -0.021 (0.55) -0.07 0.07 Electrical conductivity -0.02 (0.58) -0.10 0.05

Covariates Mantel r (p value) 2.5% llim 97.5% ulim pH 0.09 (0.18) 0.14 0.13

Partial Mantel r 0.09 (0.12) 0.03 0.16

Statistical significance, alpha = 0.05: (p-values * p <0.05, ** p <0.01, *** p <0.001)

Table 3. 3. Mantel’s test shows the relationship between pairwise dissimilarity in species composition and dissimilarity in ecological covariates in Tunga sub-basin. Species turnover was mainly influenced by geographical distance, stream characteristics (depth-width ratio, stream order), local habitat (rocks, boulders) and water chemistry (electrical conductivity and pH). ______Covariates Mantel r (p value) 2.5% llim 97.5% ulim Geographical distance ** 0.21 (0.002) 0.16 0.28 Rocks** 0.31 (0.002) 0.23 0.37 Boulders** 0.21 (0.007) 0.14 0.27 Stream order** 0.20 (0.006) 0.13 0.25 Slope 0.01 (0.36) -0.04 0.06 Depth width ratio** 0.22 (0.008) 0.08 0.27 Water temperature** 0.17 (0.009) 0.13 0.23 Electrical conductivity -0.09 (0.86) -0.14 -0.02 pH** 0.16 (0.05) 0.08 0.21

Partial Mantel r 0.19 (0.007) 0.12 0.25

Statistical significance, alpha = 0.05: (p-values * p <0.05, ** p <0.01, *** p <0.001)

The overall richness (40 species) was very high and the species composition was very different, as more than 20 unique species (Paracnthocobitis mooreh and Balitora mysorensis etc) were found in Tunga sub-basin alone, suggesting importance of stream heterogeneity (See Appendix – 3.1 for complete species list). Dissimilarity in species composition across segments with respect to a hypothetical combined pooled reference composition was higher in Bhadra than Tunga, while in case of endemic species composition was higher in Tunga than Bhadra (Appendix 3. 2 a & b). Incidence based data showed higher dissimilarity in species composition and endemic species composition in Tunga compared to Bhadra (Appendix 3. 3 a and b).

Discussion This study offers interesting insights into fish species composition (species turnover) at smaller spatial scales in two river sub-basins of the WGs, India. Geographic

59 proximity (dispersal limitation) did not result in higher species turnover, suggesting beta diversity patterns might be controlled by niche related processes (habitat complexity and biophysical factors). Among environmental factors, stream order, rocks and canopy were highly correlated with fish species composition in Bhadra sub- basin, while in Tunga sub-basin, stream substratum (gravel, rocks and boulders) determined fish species composition (Table 1). Dissimilarity in species composition across segments was higher in both the river sub-basins implying partial species turnover across sampled segments for both total and endemic species (Appendix 3.2 a & b). Dissimilarity in species composition across segments was consistently higher in Tunga (based on incidence data) than in Bhadra (Appendix 3.3 a & b) indicating higher species turnover in Tunga than in Bhadra. Stream fishes are comprised of good as well as poor dispersers. Rheophilic (with greater swimming ability) species are mostly surface dwellers (Barilius sp, Salmophasia sp, Tor sp) and seem to be highly mobile both in the main river channel as well as among all the tributaries draining the main channel (Atkore et al., 2017). However, there are headwater and habitat specialist fishes that restrict themselves to specific habitats (Balitora sp, Pethia sp, Glyptothorax sp) and therefore can be categorized as poor dispersers (Atkore et al., 2017). Water chemistry especially water temperature and dissolved oxygen (in riffle dominated habitats at headwater and mid sections of streams) seems to govern mobility of headwater fish species.

Overall, the Bhadra sub-basin was impoverished with only 20 species. The high topography and presence of a water fall could have hindered fish migration from lower elevations to the higher elevations. Many fishes were headwater specialists. On the other hand, Tunga river-sub-basin had a gentle elevational gradient with diverse stream habitats with few natural river barriers (waterfalls). Fish movement could have been easier within the main river channel as well as across different tributaries. This results in the higher observed species richness and unique species composition. Bhadra sub-basin could be treated as subset of Tunga species composition since, all the species of Bhadra (except one species i.e. Garra stenorhynchus) were also found in Tunga. Additionally, Tunga had 20 unique fish species in its habitat making it the most species rich sub-basin sampled in the study. About 19 species were common in two sub-basins (Appendix – 3.1). Thus, headwater habitat that is richer in cool temperatures, dissolved oxygen, substrate composition and canopy cover seems to be

60 vital for habitat specialist and endemic fish species. Preserving such crucial habitats is critical for sustaining endemic fish populations.

This study suffers from a few caveats. I could only use wet season data to estimate fish species turnover as adequate coverage of sampling sites was available only for this season. It is known that seasonal variation influences fish diversity and species turnover (Adams et al., 2004; Shimadzu et al., 2013; 2015). Covering multiple seasons across sampling sites would be important in future studies. The spatial coverage of this study was small, and I was able to just cover approximately 500 m elevation gradient within the Kudremukh National Park consisting mostly a headwater region. Future studies in this region should cover the lower parts of the elevational gradient as well.

This study offers insight into stream habitat variables that influence distribution of fish assemblages within a basin at high to middle elevation zones. The knowledge from this study offers insights for prioritizing stream segments with specific stream habitat characteristics for fish conservation within these basins.

References

Abraham, Robin Kurian, and Nachiket Kelkar. 2012. Do terrestrial protected areas conserve freshwater fish diversity? Results from the Western Ghats of India. Oryx No. 46 (04):544-553. doi: doi:10.1017/S0030605311000937.

Abes S. S., and Agostinho AA 2001. Spatial patterns in fish distributions and structure of the ichthyocenosis in the Água Nanci stream, upper Paraná River basin, Brazil. Hydrobiologia 445:217–227

Adams, Susan, Melvin Warren, and Wendell Haag. 2004. Spatial and temporal patterns in fish assemblages of upper coastal plain streams, Mississippi, USA. Hydrobiologia No. 528 (1):45-61. doi: 10.1007/s10750-004-1659-7.

Alfonsi, E., Benot, M-L., Fievet, V., and Alard, D. 2017. Addressing species turnover and community changes in vegetation resurvey studies. Applied Vegetation Science. 20 (2), 172-182. doi:10.1111/avsc.12258.

Al-Shami, Salman A., Jani Heino, M. R. Che Salmah, A. Abu Hassan, A. H. Suhaila, and Madziatul R. Madrus. 2013. Drivers of beta diversity of macroinvertebrate communities in tropical forest streams. Freshwater Biology No. 58 (6):1126- 1137. doi: 10.1111/fwb.12113.

Anderson, Marti J., Thomas O. Crist; Jonathan M. Chase; Mark Vellend; Brian D. Inouye; Amy L. Freestone; Nathan J. Sanders; Howard V. Cornell; Liza S. Comita; Kendi F. Davies; Susan P. Harrison; Nathan J. B. Kraft; James C. Stegen; Nathan G. Swenson. 2011. Navigating the multiple meanings of β diversity: a roadmap for the practicing ecologist. Ecology Letters No. 14 (1):19-28. doi: 10.1111/j.1461-0248.2010.01552.x.

APHA. 2005. Standard Methods for Examination of Water and. Wastewater, American Public Health Association WWA,. Washington, D.C. 23rd ed. Pp. 1504

Arthington, Angela H., Robert J. Naiman, Michael E. McClain, and Christer Nilsson. 2010. Preserving the biodiversity and ecological services of rivers: new challenges and research opportunities. Freshwater Biology No. 55 (1):1-16. doi: 10.1111/j.1365-2427.2009.02340.x.

Arunachalam, M. 2000. Assemblage structure of stream fishes in the Western Ghats (India). Hydrobiologia No. 430 (1-3):1-31. doi: 10.1023/A:1004080829388.

Atkore, V., N. Kelkar., and J. Krishnaswamy. 2017. Assessing the recovery of fish assemblages dowstream of hydrological barriers in India's Western Ghats. River Research and Application. Doi:10.1002/rra.3163.

Baselga, Andrés. 2010. Partitioning the turnover and nestedness components of beta diversity. Global Ecology and Biogeography No. 19 (1). doi: 10.1111/j.1466- 8238.2009.00490.x.

Baselga, Andrés. 2012. The relationship between species replacement, dissimilarity derived from nestedness, and nestedness. Global Ecology and Biogeography No. 21 (12). doi: 10.1111/j.1466-8238.2011.00756.x.

Baselga, Andrés. 2013. Multiple site dissimilarity quantifies compositional heterogeneity among several sites, while average pairwise dissimilarity may be misleading. Ecography No. 36 (2). doi: 10.1111/j.1600-0587.2012.00124.x.

Bertuzzo, E., R. Muneepeerakul, H. J. Lynch, W. F. Fagan, I. Rodriguez- Iturbe and A. Rinaldo. 2009. On the geographic range of freshwater fish in river basins. Water Resource Research No. 45 (W11420). doi: 10.1029/2009WR007997.

Bhat, A. 2003. Diversity and Composition of Freshwater Fishes in River Systems of Central Western Ghats, India. Environmental Biology of Fishes No. 68 (1):25- 38. doi: 10.1023/a:1026017119070.

Bhat, A., and A. E. Magurran. 2007. Does disturbance affect the structure of tropical fish assemblages? A test using null models. Journal of Fish Biology No. 70 (2):623-629. doi: 10.1111/j.1095-8649.2006.01252.x

Borcard, D., P. Legendre and Pierre Drapeau. 1992. Partialling out the spatial component of ecological variation. Ecology No. 73 (3):1045-1055.

Boyero, L. 2003. Multiscale patterns of spatial variation in stream macroinvertebrate communities. Ecological Research No. 18:365-379.

Boyero, Luz, Richard G. Pearson, David Dudgeon, Manuel A. S. Graça, Mark O. Gessner, Ricardo J. Albariño, Verónica Ferreira, Catherine M. Yule, Andrew J. Boulton, Muthukumarasamy Arunachalam, Marcos Callisto, Eric Chauvet, Alonso Ramírez, Julián Chará, Marcelo S. Moretti, José F. Gonçalves, Julie E. Helson, Ana M. Chará-Serna, Andrea C. Encalada, Judy N. Davies, Sylvain Lamothe, Aydeè Cornejo, Aggie O. Y. Li, Leonardo M. Buria, Verónica D. Villanueva, María C. Zúñiga, and Catherine M. Pringle. 2011. Global distribution of a key trophic guild contrasts with common latitudinal diversity patterns. Ecology No. 92 (9):1839-1848. doi: 10.1890/10-2244.1.

Burridge, Christopher P., Dave Craw, Daniel C. Jack, Tania M. King, and Jonathan M. Waters. 2008. Does fish ecology predict dispersal across a river drainage

divide? Evolution No. 62 (6):1484-1499. doi: 10.1111/j.1558- 5646.2008.00377.x.

Chang M H, Yao-Sung Lin and Lin-Chung Chung 1999. Effect of dams of fish assemblages of the Tachia Rriver, Taiwan. Acta Zoologica Taiwanica No. 10:77-90.

Crawley, M. J. 2007. The R Book. The Atrium Southern Gate, Chichester, West Sussex PO 19 8SQ, England. : John & Wiley Sons, Ltd. Pp. 877.

Dallas, Helen. 2008. Water temperature and riverine ecosystems: An overview of knowledge and approaches for assessing biotic responses with special reference to South Africa. Water SA No. 34 (3).

Daniels, R.J.R. 2002. Freshwater fishes of Peninsular India. Hyderabad: University Press India (Pvt) Ltd. Pp.320.

Das, Arundhati, Jagdish Krishnaswamy, Kamaljit S. Bawa, M. C. Kiran, V. Srinivas, N. Samba Kumar, and K. Ullas Karanth. 2006. Prioritisation of conservation areas in the Western Ghats, India. Biological Conservation No. 133 (1):16-31. doi: http://dx.doi.org/10.1016/j.biocon.2006.05.023.

Dornelas, Maria. 2010. Disturbance and change in biodiversity. Philosophical Transactions of the Royal Society B: Biological Sciences No. 365 (1558):3719-3727. doi: 10.1098/rstb.2010.0295.

Dudgeon, D. 2011. Asian river fishes in the Anthropocene: threats and conservation challenges in an era of rapid environmental change. Journal of Fish Biology No. 79 (6):1487-1524. doi: 10.1111/j.1095-8649.2011.03086.x.

Dudgeon, David. 2000. Large-scale hydrological changes in Tropical Asia: Prospects for riverine biodiversity. BioScience No. 50:9.

Edds, David R. 1993. Fish Assemblage Structure and Environmental Correlates in Nepal's Gandaki River. Copeia No. 1993 (1):48-60. doi: 10.2307/1446294.

Fialho AP, Oliveira LG, Tejerina-Garro FL, and Mérona B 2008. Fish-habitat relationship in a tropical river under anthropogenic influences. Hydrobiologia 598:315–324.

Gaston, Kevin J. 2000. Global patterns in biodiversity. Nature No. 405 (6783):220- 227.

Gering, Jon C., Thomas O. Crist, and Joseph A. Veech. 2003. Additive Partitioning of Species Diversity across Multiple Spatial Scales: Implications for Regional Conservation of Biodiversity. Conservation Biology No. 17 (2):488-499. doi: 10.1046/j.1523-1739.2003.01465.x. Gido, K., and Donald A. Jackson. 2010. Community Ecology of Stream Fishes: Concepts, Approaches, and Techniques. K. Gido, Donald A. Jackson (ed). Bethseda, Maryland, USA: American Fisheries Society Sympposium 73. Pp. 665. Gillette, D., J. S Tiemann, D R. Edds and M. L. Wildhaber. 2005. Spatiotemporal patterns of fish assemblage structure in a Rriver impounded by low-head dams. Copeia No. 3:539-549.

Gorman, Owen T., and James R. Karr. 1978. Habitat Structure and Stream Fish Communities. Ecology No. 59 (3):507-515. doi: 10.2307/1936581.

Goslee, S and, and Dean Urban. 2007. The ecodist package for dissimilarity-based analysis of ecological data. Journal of Statistical softwarre No. 22 (7):1-19.

Hauer, F.R and, and G. Lamberti. 2007. Methods in stream ecology. F.R and G. Lamberti Hauer (ed). Vol. 2nd: Academic Press.Pp. 896.

Heino, Jani, Janne Soininen, Janne Alahuhta, Jyrki Lappalainen, and Risto Virtanen. 2015. A comparative analysis of metacommunity types in the freshwater realm. Ecology and Evolution No. 5 (7):1525-1537. doi: 10.1002/ece3.1460.

Heller, E., J.M. Rhemtulla, S. Lele, M. Kalacska, S. Badiger, R. Sengupta and, and N. Ramankutty. 2012. Mapping crop types, irrigated areas and cropping intensities in heterogenous landscapes of southern India using multi-temporal medium resolution imagery: Implications for assessing water use in agriculture. Photogrammetric engineering and Remote Sensing No. 78 (8):815-827.

Hugueny, B., T. Oberdorff., and P. A. Tedescco. 2010. Community ecology of river fishes: a large-scale perspective. In Community Ecology of Stream Fishes: Concepts, Approaches and Techniques B. Gido Keith, and D. A. Jackson (ed). Pp 1-34. Bethesda, Maryland, USA: American Fisheries Society.

Jackson, Donald A., Pedro R. Peres-Neto, and Julian D. Olden. 2001. What controls who is where in freshwater fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries and Aquatic Sciences No. 58 (1):157-170. doi: 10.1139/f00-239.

Jardine Timothy D., Nicholas R. Bond, Michele A. Burford, Mark J. Kennard, Douglas P. Ward, Peter Bayliss,Peter M. Davies, Michael M. Douglas,Stephen K. Hamilton, John M. Melack, Robert J. Naiman, Neil E. Pettit,Bradley J. Pusey, Danielle M. Warfe, and Stuart E. Bunn. 2015. Does flood rhythm drive ecosystem responses in tropical riverscapes? Ecology No. 96 (3):684–692.

Jayaram, K. C. 2010. The Frehswater Fishes of Indian Region. Delhi: Narendra Publishing House.Pp. 551.

Johnson, J.A. , and M. Arunachalam. 2009. Diversity, distribution and assemblage structure of fishes in streams of southern Western Ghats, India. Journal of Threatened Taxa No. 1 (10):507-513.

Johnson, J. A., and M. Arunachalam. 2010. Habitat use of fishes in streams of Kalakad Mundanthurai Tiger Reserve, India. International Journal of Ecology and Development No. 17 (F 10):34-47.

Jost, L. 2007. Partitioning diversity into independent alpha and beta components. Ecology No. 88:2427-2439.

Kasturirangan, K. 2013. Report of the High Level Working Group on Western Ghats. Ministry of Environment and Forests, Government of India.

Koleff, P., K. J. Gaston and J. J. Lennon. 2003. Measuring beta diversity for presence- absence data. Journal of Animal Ecology No. 72:367-382.

Krishnaswamy, Jagdish, Kamaljit S. Bawa, K. N. Ganeshaiah, and M. C. Kiran. 2009. Quantifying and mapping biodiversity and ecosystem services: Utility of a multi-season NDVI based Mahalanobis distance surrogate. Remote Sensing of Environment No. 113 (4):857-867. doi: http://dx.doi.org/10.1016/j.rse.2008.12.011.

Lande, R. 1996. Statistics and partitioning of species diversity, and similarity among multiple communities. Oikos No. 76:5–13.

Leprieur, Fabien, Julian D. Olden, Sovan Lek, and Sébastien Brosse. 2009. Contrasting patterns and mechanisms of spatial turnover for native and exotic freshwater fish in Europe. Journal of Biogeography No. 36 (10):1899-1912. doi: 10.1111/j.1365-2699.2009.02107.x.

Magurran, A. E. 2004. Measuring biological diversity. Oxford: Blackwell Science.Pp. 256.

Magurran, Anne E., and A. T. Phillip Dawn. 2001. Implications of Species Loss in Freshwater Fish Assemblages. Ecography No. 24 (6):645-650.

Malkhede, S. K. 2003. Working Plan for Belgaum Forest Division (2002-2003 to 2011-2012). Belgaum, Karnataka state, India: Belgaum Forest Division, Belgaum.

Matthews, M. J. 1998. Patterns in freshwater fish ecology. Massachusetts, USA.: Kluwer Academic publisher.Pp.670.

McGill, Brian J., Maria Dornelas, Nicholas J. Gotelli, and Anne E. Magurran. 2015. Fifteen forms of biodiversity trend in the Anthropocene. Trends in Ecology and Evolution No. 30 (2):104-113. doi: http://dx.doi.org/10.1016/j.tree.2014.11.006.

McKnight, Megan Wilde. 2007. Broad-scale patterns and determinants of beta diversity. A PhD thesis submitted to the University of North Carolina, Chapel Hill.

Molur, S., K. G. Smith, B. A. Daniel, and W. R. T. Darwall. 2011. The Status and distribution of freshwater biodiversity in the Western Ghats, India. Cambridge, UK and Gland, Switzerland: IUCN, and Coimbatore, India: Zoo Outreach Organisation.: IUCN.

Ott, R. Lyman, and Michael T. Longneck. 2010. An introduction to statistical methods and data analysis. 6 ed. Texas A and M University.

Pegg, M. A. and R. M. Taylor. 2007. Fish species diversity among spatial scales of altered temperate rivers. Journal of Biogeography No. 34:549-558. doi: 10.1111/j.1365-2699.2006.01624.x.

Philippi, Thomas E., Philip M. Dixon, and Barbara E. Taylor. 1998. Detecting Trends in Species Composition. Ecological Applications No. 8 (2):300-308. doi: 10.2307/2641069.

Poff, L. M. and Allan J. D. 1995. Functional organization of stream ﬁsh assemblages in relation to hydrologic variability. Ecology No. 76:606-627.

Pringle, C. M. 2001. Hydrologic connectivity and the management of biological reserves: a global perspective. Ecological Application No. 11 (4):981-998.

Robert, C. Szava-Kovats, and Pärtel Meelis. 2014. Biodiversity Patterns along Ecological Gradients: Unifying β-Diversity Indices. PLoS ONE. doi: 10.1371/journal.pone.0110485.

Sakaris, P. C. 2013. A review of the effects of hydrologic alteration on fisheries and biodiversity and the management and conservation of natural resources in regulated river systems. Paul M. Bradley (ed). Current perspectives in contaminant hydrology and water resource sustainability.

Schlosser, Isaac J. 1991. Stream Fish Ecology: A Landscape Perspective. BioScience No. 41 (10):704-712. doi: 10.2307/1311765.

Shimadzu, Hideyasu, Maria Dornelas, Peter A. Henderson, and Anne E. Magurran. 2013. Diversity is maintained by seasonal variation in species abundance. BMC Biology No. 11 (1):98. doi: 10.1186/1741-7007-11-98.

Shimadzu, Hideyasu, Maria Dornelas, and Anne E. Magurran. 2015. Measuring temporal turnover in ecological communities. Methods in Ecology and Evolution No. 6 (12):1384-1394. doi: 10.1111/2041-210X.12438.

Soininen, Janne. 2010. Species Turnover along Abiotic and Biotic Gradients: Patterns in Space Equal Patterns in Time? BioScience No. 60 (6):433-439. doi: 10.1525/bio.2010.60.6.7.

Soininen, Janne, Robert McDonald, and Helmut Hillebrand. 2007. The distance decay of similarity in ecological communities. Ecography No. 30 (1):3-12. doi: 10.1111/j.0906-7590.2007.04817.x.

Tondato, K. K. and Y. R. Suarez. 2010. Temporal changes in fish species composition of headwater streams of the upper Praguay and Parana basins, Brazil. Acta Scientiarum Biologcal Sciences No. 32 (3). doi: DOI: 10.4025/actascibiolsci.v32i3.6160.

Trivedi, R.K., and P.K. Goel. 1986. Chemical and Biological methods for water pollution studies: Environmental Publications, India.Pp. 250.

Tuomisto, Hanna. 2010. A diversity of beta diversities: straightening up a concept gone awry. Part 1. Defining beta diversity as a function of alpha and gamma diversity. Ecography No. 33 (1):2-22. doi: 10.1111/j.1600-0587.2009.05880.x.

Unmack, P T. 2001. Biogeography of Australian Freshwater fishes. Journal of Biogeography No. 28 (9):1053-1089.

Vasconcelos, L. P., D. C. Alves, and L. C. Gomes. 2014. Fish reproductive guilds downstream of dams. Journal of Fish Biology No. 85 (5):1489-1506. doi: 10.1111/jfb.12501.

Vasudevan, K., Ajith Kumar, and Ravi Chellam. 2006. Species turnover: the case of stream amphibians of rainforests in the Western Ghats, southern India. In Marine, Freshwater, and Wetlands Biodiversity Conservation. David L. Hawksworth and AlanT Bull (ed). Springer Netherlands. Pp. 147-157.

Veech, Joseph A. and Thomas O. Crist. 2010. Diversity partitioning without statistical independence of alpha and beta. Ecology No. 91:1964–1969. doi: http://dx.doi.org/10.1890/08-1727.1.

Vorosmarty, C. J., P. B. McIntyre, M. O. Gessner, D. Dudgeon, A. Prusevich, P. Green, S. Glidden, S. E. Bunn, C. A. Sullivan, C. Reidy Liermann, and P. M. Davies. 2010. Global threats to human water security and river biodiversity. Nature No. 467 (7315):555-561. doi: http://www.nature.com/nature/journal/v467/n7315/abs/nature09440.html#supp lementary-information.

Whittaker, R. H. 1960. Vegetation of Siskiyou Mountains, Oregon and California. Ecological Monograph No. 30:279-338.

Wiersma, Yolanda F., and Dean L. Urban. 2005. Beta Diversity and Nature Reserve System Design in the Yukon, Canada. Conservation Biology No. 19 (4):1262- 1272. doi: 10.1111/j.1523-1739.2005.00099.x. Wikramnayake, E. D. 1990. Ecomorphology and biogeography of a tropical stream fish assemblage: evolution of assemblage structure. Ecology No. 71 (5):1756- 1764.

Wilson, M.V. and Shmida, A. 1984. Measuring beta diversity with presence–absence data. Journal of Ecology No. 72:1055–1064.

Wurts, W.A. , and M. Durborow Robert. 1992. Interactions of pH, carbon dioxide, alkalinity and hardness in fish ponds. Southern Regional Aquaculture Center, USA (464):1-4.

Yap, S. Y. 2002. On the distributional patterns of Southeast-East Asian freshwater fish and their history. Journal of Biogeography No. 29 (9):1187-1199. doi: 10.1046/j.1365-2699.2002.00771.x.

Chapter 4 River barriers and environmental variables influence fish guild responses in the Western Ghats of India

Publication Atkore V., N. Kelkar., S. Badiger., K. Shanker and J. Krishnaswamy. 2017. River barriers and environmental variables influence fish guild responses in the Western Ghats of India. In Prep.

River barriers and environmental variables influence fish guild responses in the Western Ghats of India

Introduction Freshwater is vital to sustain biodiversity as well as human well-being (Baron, 2003; Dudgeon et al., 2006; Vorosmarty et al., 2010), but freshwater ecosystems are among the most affected by anthropogenic impacts in the world (Dudgeon, 2010; Strayer, 2010; Vorosmarty et al., 2010) and a high proportion of riverine biodiversity is endangered (Darwall et al., 2008; Dudgeon, 2000; Moreno and Rodriguez, 2010; Nel et al., 2007). As tropical rivers have been altered by various anthropogenic interventions including dams, barrages, and other river regulation schemes to meet water and energy needs, the impacts of flow regulation and water abstraction on freshwater biodiversity and ecosystem services have been generally neglected (Bunn and Arthington, 2002; Moreno and Rodriguez, 2010; Poff and Zimmerman, 2010). As a result, the highly diverse taxon of freshwater fishes (>5000 species worldwide) is listed as the most threatened vertebrate group by the International Union for Conservation of Nature (Reid et al., 2013).

Man-made barriers disrupt the longitudinal and lateral connectivity of river systems (Pringle, 2001; Vannote et al., 1980), often with highly negative effects on fish communities (Konar et al., 2013; Malmqvist and Rundle, 2002; Poff and Allan, 1995). Dams have both immediate as well as long-term effects on life histories of fishes (Vasconcelos et al., 2014a, b). Though dams and reservoirs might benefit some generalist (and invasive) fish species (Chu et al., 2015), overall they negatively affect the diversity, abundance, life-history traits, and breeding success of many specialist and endemic fish species (Dudgeon, 2000; Pringle, 2001; Mims and Olden, 2012; 2013; Reid et al., 2013). For instance, dewatering by a small dam in Costa Rica in the downstream section affected the life history strategies of fishes such as cichlids, with complex reproductive requirements (Anderson et al., 2006). A study in the Colorado River found that regions with high densities of dams benefited invasive and exotic species with equilibrium life history strategies over native fish species with opportunistic / periodic breeding and growth patterns (Pool et al., 2010). Similarly, hydropeaking operations downstream of the Itutinga dam led to the disappearance of

72 insectivorous fish communities in the Grande river basin in Brazil (Gandini et al., 2014). Apart from hydrological and environmental variables, natural factors such as basin geomorphology, biotic interactions, land use history, as well as anthropogenic disturbances such as water abstraction, also influence fish communities (Gilliam et al., 1993; Jackson et al., 2001b). Local human-induced disturbances from illegal fishing, pollution, and substrate mining pose further threats to fish feeding and breeding habitats (Daniels, 2002). Such disturbances can either exert these effects independently, or flow barriers can aggravate their effects. As a result, it often becomes difficult to separate local anthropogenic effects from larger basin-scale effects due to modifications of water quantity and quality. Knowledge is still fairly limited about the impacts of multi-scale threats on freshwater fish diversity in tropical Asian river systems (Bhat, 2002).

Available studies on freshwater fish diversity describe the functional responses of fish guilds to riverine habitat gradients or stream characteristics, but do not adequately quantify the effects of environmental covariates and local disturbances in relation to flow regulation (De Silva et al., 1979; Bhat, 2004; Weliange and Amarsinghe, 2007; Johnson and Arunachalam, 2012; Chakravarty and Homechaudhuri, 2013). But, few recent studies have quantified and modelled the functional guilds of fishes in relation to flow regime and habitat characteristics (Oliveira et al., 2012; Macnaughton et al., 2016). In India, despite a high diversity of freshwater fish fauna, studies on factors driving persistence of various fish guilds in hydrologically altered and human- modified river systems are relatively few (Jayaram, 2010; Raghavan et al., 2016). Guild-based approaches can provide an intuitive understanding of the functional responses of fish communities to alterations (Villager, 2008) and could help in developing a broad understanding of the impact of river modification on fishes (Mouillot et al., 2013). Earlier studies from the global biodiversity hotspot of the Western Ghats (WG) in India showed that fish communities partitioned habitat space based on their eco-morphology and in relation to stream substrate (Bhat, 2005), higher species richness occurred at lower elevations (Raghavan et al., 2008), and that regulated rivers differed strongly from non-regulated rivers in fish assemblage structure (Bhat and Magurran, 2007). Despite a growing body of work on fish taxonomy and community ecology in India, relatively few studies have explicitly compared guild-wise species richness in regulated and non-regulated river basins

(Kundu et al., 2014). I adopted a guild-based approach to assess ecological responses of fish species richness of three guilds (based on position in water column) to river barriers, water chemistry, and local water abstraction in the central Western Ghats (WG’s) of India. I defined river barriers as “anthropogenic modifications of flows by dams or barrages to alter the river’s flow regime”. River barriers are mainly major and minor barrier. Minor barriers consist of small check dams with height of < 1.2 m, width < 50 m and an impoundment area of < 10 sq.km while major barriers include check dams with highest > 1.5 – 2 m, width > 50 -150m and an impoundment area > 11 sq.km (Atkore et al., 2017). Based on comparisons of regulated and non-regulated sub-basins (called RSB and NRSB), the study addressed three questions: 1) how do fish guild richness and species composition vary across regulated and non-regulated sub-basins? 2) how do environmental covariates and local anthropogenic disturbance (water abstraction) differ between RSB and NRSB? and 3) how do environmental covariates vary with guild-wise fish species richness in RSB and NRSB? Based on predictions of previous studies (Brasher, 2003; Merz, 2013; Macnaughton et al., 2016), I hypothesized that river segments in regulated sub-basins will have higher temperature and poorer water quality than river segments in non-regulated basins. I discuss the results in light of the potential impacts of river barriers and local disturbance on three water column position-based guilds. I also discuss the choice of this guild-based classification for studies on freshwater fishes in human modified tropical river systems. Finally, I outline the implications of our findings for mitigating effects of flow regulation, monitoring river habitats, and developing guidelines for ecological flow regimes in the region.

Materials and Methods Study Site The study was conducted in the states of Karnataka and Goa in the central WG’s region of India from 2011-2014 (Figure 1). The WG mountain range along the western coast of India is a distinct zoogeographical subdivision of Peninsular India (Bhimachar, 1945) and a global biodiversity hotspot and a world heritage site (Cincotta et al., 2000; Das et al., 2006), with some of the highest human population densities in the world. I conducted sampling across two regulated sub-basins (RSB,

Malaprabha and Mhadei), and two non-regulated sub-basins (NRSB, Bhadra and Tunga). The Malaprabha and Mhadei (aka Mahadayi/Mandovi) originate from the same hill range at 760 m elevation bordering the states of Goa and Karnataka. The headwater catchments of Malaprabha and Mhadei are dominated by tropical moist evergreen forest and the downstream plains are mostly covered by agriculture and some scrub forest. These sub-basins receive heavy southwest monsoon rainfall from June to September. In Malaprabha, the annual rainfall varies from 2000 to 3500 mm and average annual stream discharge is 1944 million cubic metres (MCM) yr-1. Average annual rainfall recorded for the Mhadei sub-basin was 3955 mm and average annual stream discharge was 3447 MCM yr-1 (Ibrahampurkar, 2012). There are numerous barrages and check-dams built on both rivers to store dry season flows from the headwater catchments for irrigation (Ibrahampurkar, 2012). An inter-basin water transfer project is underway on two headwater tributaries of Mhadei sub-basins to divert 7.56 TMC (thousand million cubic feet) water annually to provide drinking water for some towns and villages in Karnataka State. In the last two decades, the cropping pattern in the Malaprabha catchment has changed from rainfed crops such as millets and pulses to water-intensive crops such as sugarcane, vegetables, and oilseeds. As a result, water extraction from the river and groundwater has intensified (Heller et al., 2012), drastically affecting stream flows during both the monsoon and dry season. Municipal wastewater and fertilizer runoffs from agricultural fields have seriously lowered river fish productivity, according to local communities in the area. Other local disturbances to stream biota in the area are from fishing and sand-boulder mining.

The non-regulated sub-basins of Tunga and Bhadra are spread across the Chikkamgaluru, Dakshina Kannada, and Udupi districts of Karnataka. The headwater streams originate at approx. 1160 m elevation in the Kudremukh National Park (KNP). The region receives more than 6000 mm of rainfall annually (Krishnaswamy et al., 2006). These sub-basins have evergreen forests in the headwaters, while downstream reaches have interspersed paddy fields and coffee plantations (Appendix 4.1 -S1A Table).

Figure 4. 1. Sampling locations in the four sub-basins in the central Western Ghats, India (1 = Malaprabha 2 = Mhadei 3 = Tunga 4 = Bhadra).

Hydrological time-series for regulated river sub-basins (RSB) Data on river discharge (cumecs) were obtained for Malaprabha and Mhadei for the period from 1979 to 2013 from the Water Resource Development Organization (WRDO), Karnataka and the Department of Water Resources, Goa. Streamflow data from gauging stations located at Collem and Ganjem (for Mhadei river), and Khanapur (for Malaprabha river) were used. Similar datasets could not be obtained for the non-regulated Bhadra and Tunga sub-basins due to absence of stream gauging stations in this part of the study area. Daily rainfall time series data (in mm) were extracted from gridded APHRODITE datasets (spatial resolution of 0.25 x 0.25 degrees (Yatagai et al., 2012) to estimate basin-averaged annual rainfall from 1979 to 2013.

Fish sampling Fish communities were sampled and corresponding ecological variables were collected from the segments of approximately 150 m length across the 4 sub-basins. The chosen river segments located in both regulated and non-regulated sub-basins had similar elevation ranges (600 m to 930 m), land-use, stream order (2-6) and habitat characteristics (pool, run, riffles) to control for the effects of these variables on species turnover to the extent possible. At each of these river segments, fish sampling was conducted with cast-nets of 0.5 cm x 0.5 cm and 1.1 cm x 1.1 cm mesh-sizes. I relied on using cast nets over other fishing methods because they are known to provide the best coverage and capture of different fish guilds (irrespective of position in the water column) in rivers of the WG’s. Further, castnets are a non-destructive fishing technique with the lowest fish mortality (Abraham and Kelkar, 2012; Ahmad et al., 2013) among competing and more intensive methods. Ease of operation, high sampling efficiency and portability enable cast nets to provide better access in sampling locations and habitats in relatively remote river reaches. The overall cast-net effort yielded 79 fish species, which represented nearly 75.23% of the total pool (105 species) known from the study area (Rema Devi et al., 2013), from which we deemed the overall effort as adequate. Sampling was continued until saturation in species accumulation with increasing cast-net effort was observed at each segment (Abraham and Kelkar, 2012). Each replicate thus included the total occurrence data from all casts (10 to 26) conducted over approximately 120 minutes at each river segment. In each segment I conducted cast-netting strategically and based on initial checks on the

77 feasibility of sampling run, riffle, and pool habitats. At some sections of river habitat, visual methods were used (Arunachalam, 2000; Deacon et al., 2015) to complement cast-net effort in recording species occurrence. At each river segment, fishes were caught, identified, measured (total length in cm) and released live into the water. Only 3-5 individuals of unidentified fish species were collected and preserved in 4% formaldehyde solution for further identification in the lab. Standard field guides (Daniels, 2002; Jayaram, 2010) were used for species-level identification and validation (Eschmeyer, 2015). Any evidence of disease or deformity in the fish species sampled was recorded in each segment.

Environmental and water quality variables such as river depth, temperature, dissolved oxygen, calcium hardness, total dissolved solids, canopy cover, river substrates (%) were measured for all segments (Appendix S1B Table). I also recorded the level of disturbance (low, medium, or high) based on observed activities deemed harmful to fish species (water abstraction, fishing, substrate-mining, etc). The intensity of these activities was ranked through direct visual observations.

Guild-based classification of fish species I grouped all sampled fish species according to six classification schemes (see analysis methods and Table 4. S2), based on intensive field surveys and a review of published (journal articles, field guides) and unpublished (reports, monographs) literature. Fish species were classified based on depth or water column use i.e. surface dwellers, mid-column dwellers and bottom dwellers (Lowe-McConnell, 1975, 1987; Arunachalam, 2000; Bhat, 2002), feeding preferences, (phytophagous, heterotrophic, and omnivorous) (Arunachalam, 2000; Bhat, 2002; Weliange and Amarsinghe, 2007; Johnson and Arunachalam, 2012; Froese and Pauly, 2016;), flow-responses (eurytopic, limnophilic, and rheophilic) (Aarts and Nienhuis, 2003; Chakrabarty and Homechaudhuri et al., 2013), sensitivity to disturbance (Daniels, 2002) reproductive strategies (lithopelagophils, lithophils, pelagophils, phytolithophils, phytophils and psammophils), and life-history cues (equilibrium, opportunistic, periodic/seasonal and intermediate strategist) (Welcomme, 1985; Winemiller et al., 2006; 2008), and endemicity to the WG’s region (Daniels, 2002; Dahanukar and Raghavan, 2013).

Data Analysis Choosing representative guilds for further analyses Non-metric Multidimensional Scaling (NMDS) is an ordination method that uses dissimilarities calculated from community data collected across sites. This assists in segregating sites with similar assemblages that tend to be closer to each other than sites with more dissimilar assemblages (Borcard et al., 2011). Instead of conventional site-by-species ordination, a trait-by-species ordination with NMDS was used to identify how species clustered together based on the similarity of their classification into different guilds (based on life-history traits related to feeding, reproduction, movement etc.). A presence-absence matrix was prepared for all species (in columns) and all traits (in rows) to run the NMDS analysis until it attained a stable configuration (Borcard et al., 2011). Based on the analysis, I found that guild classification based on position of fishes in the water column (surface-, mid-column-, and bottom-dwelling fishes) was the most stable and representative of other life- history based classifications that were correlated with fish guilds. The richness of each water column position-based guild was defined as the response variable for further analyses. The guild-based analysis also helped overcome any biases resulting from unknown heterogeneity in taxonomic resolution or systematic biogeographic variations in fish species richness across the RSB and NRSB.

Trend analyses of rainfall and discharge data Time-series analyses were performed to estimate trends in discharge (cumecs, minimum and maximum, yr-1) and annual rainfall (mm, yr-1) for stations in regulated sub-basins. Sen’s slopes were calculated and univariate Mann-Kendall significance tests performed to assess monotonicity of trends in the average, maximum, and minimum values of annual discharge from 1979 to 2013 (Yue et al., 2002) using the R packages ‘trend’ and ‘EcoHydroLogy’ (Fuka et al., 2014; Pohlert, 2016).

Variation in fish species composition across RSB and NRSB Non-metrical multidimensional scaling was used to check whether species composition differed across the four sub-basins. NMDS ordination was unconstrained by environmental variables and driven only by species composition (Rowe, 2007). Environmental variables were subsequently fitted on the NMDS ordination axes to

79 examine which ones correlated with dissimilarities in fish community composition, using the Bray-Curtis measure of dissimilarity (Borcard et al., 2011). These analyses were conducted using the ‘vegan’ package in the R 3.2.1 software (R core Team, 2013).

Effects of environmental variables on guild-wise fish species richness in RSB and NRSB To test my hypothesis on the effects of flow regulation and anthropogenic disturbance (water extraction) on environmental variables, I visually compared and statistically tested for significance of differences between the values of environmental variables and water quality across 1) regulated and non-regulated basins, and 2) local water abstraction and pollution.

Principal Component Analysis (PCA) was used to identify correlated variables and to select variables for further regression analyses. Generalized linear regression models (GLMs) with Poisson errors were used to explore the influences of environmental and water quality variables on guild-wise fish species richness separately for RSB and NRSB. Data from the sub-basins were not analyzed together to account for potential variation in fish species detectability and any other sampling effects. I also used generalized linear mixed-effects models with ‘basin’ as the random effect variable, but found that these models provided estimates very similar to GLMs, and with inconsequential random effects. Variables that had common positive or negative effects on fish guilds across the different sub-basins were later identified. The model fit based on McFadden’s Pseudo-R2 and Akaike Information Criterion (AIC) was used to assess and compare GLMs. The models with the highest fit and lowest AIC were chosen as the best models. GLM analyses were performed in the R core Team (2013).

Results River discharge trends in RSB Total rainfall (mm yr-1) showed insignificant change in the Malaprabha and Mhadei sub-basins (Figure 4.2, Table 4. 1). However, a strong negative trend was noted in annual discharge (cumecs) for the period 1979 to 2013 across the gauging stations of the Malaprabha and Mhadei sub-basins (Figure 4. 2, Table 4. 1), highlighting the

80 impact of river regulation.

Figure 4. 2. Time-series of maximum (left column) and minimum (right column) values of discharge (cumecs, i.e. m3/s) and total rainfall (mm/yr) for three hydrological gauging stations in the Mhadei and Malaprabha sub-basin. (a) Mhadei- Collem –maximum discharge (b) Mhadei –Collem- minimum discharge (c) Mhadei- Ganjem –Maximum discharge (d) Mhadei-Ganjem –Minimum discharge (e) Malaprabha – Khanapur –Maximum discharge (f) Malaprabha-Minimum discharge.

Table 4. 1. The Mann-Kendall (M-K) test of significance for total rainfall (mm/yr), and maximum and minimum discharge (m3/s). A sharp declining monotonic trend in river discharge is noted from 1979 to 2013 despite no significant trends in rainfall time- series, which is likely due to river flow regulation and abstraction of water for irrigation. Sub-basin Station M-K statistic (S) Kendall’s Tau p-value Total rainfall (mm/yr) Mhadei Collem 12 0.03 0.83 Ganjem -17 -0.04 0.76 Malaprabha Khanapur -9 -0.02 0.88 Bhadra Kudremukh -21 -0.05 0.70 Tunga Kerekatte 23 0.05 0.67 Discharge (m3/s) Mhadei Collem (max) -147 -0.25 0.04* Collem (min) -192 -0.32 0.003*** Ganjem (max) -157 -0.28 0.02* Ganjem (min) -300 -0.504 <0.0001*** Malaprabha Khanapur (max) -22 -0.058 0.68 Khanapur (min) -58 -0.153 0.03* Bhadra Kudremukh NA NA NA Tunga Kerekatte NA NA NA

Statistical significance, alpha = 0.05: (p-values * p <0.05, ** p <0.01, *** p <0.001)

Fish species richness in different habitat guilds About 12,840 individuals of fish comprising 79 species belonging to 7 orders and 15 families in the 4 sub-basins were sampled from the period 2011 to 2014. Cypriniformes was the most dominant order, with 55 species, (44 endemic species to the WG region) and 67 % of the collected individuals. The most consistent and representative guild classification was found to be based on position of fishes in the water column, by the ordination analysis (Figure 3). Bottom-dwelling fishes (BD) consisted of 5 orders, 9 families and 41 species; mid-column fishes (MCD) had 2 orders, 3 families and 28 species; and surface-dwelling fishes (SD) had 2 orders, 3 families and 10 species (Appendix S2 Table contains details of traits associated with BD, MCD, and SD).

Figure 4. 3. NMDS ordination of the first and second axes shows guild grouping based on fish guild richness. Three clusters shows correlated guild classifications for fish species across the four sub-basins. (Guilds: SD – Surface dwelling, MCD – Mid-column dwelling and BD – Bottom dwelling, EQ – Equilibrium, OPP – Opportunistic, PER – Periodic, IMD – Intermediate, EURY – Eurytopic, LIM-Limnophilic, RHEO- Rheophilic, PLP – Polyphils, LITHO – Lithophils, PHP – Phytophils, PHPG – Phytopelagophils, LPP- Lithopelagophils).

Fish species composition across RSB and NRSB Species composition was clearly different across RSB and NRSB (Figure 4. 4). The RSBs (Mhadei and Malaprabha) also differed from each other in species composition. Exploratory analyses indicated that most rare and endemic species i.e. 72% (32 out of 44 WG endemic species) were restricted to non-regulated river basins. SD guild richness was higher in NRSB than RSB (df = 52.41, t =3.43, p =0.001), whereas species richness was higher for MCD (df = 27.38, t =-3.39, p =0.002) and BD (df = 27.00, t =-3.98, p =0.0004) in RSB than NRSB (Figure 4.5).

Figure 4. 4. NMDS ordination plot shows fish species composition across the four sub-basins (open circles indicate species). Non-regulated (Bhadra, Tunga) and regulated basins (Malaprabha, Mhadei) differed distinctly in species composition. The Mhadei and Malaprabha also clearly differed from each other in species composition.

Figure 4. 5. Guild-wise fish species richness in non-regulated and regulated sub- basins. Surface-dwelling guild richness (a) was higher in non-regulated sub-basin, while mid-column (b) and bottom-dwelling richness (c) was lower in non-regulated as compared to regulated sub-basins.

Figure 4. 6: Differences in selected water quality variables between regulated and non-regulated river sub-basins (NR=Non-regulated sub-basin, R=Regulated sub- basin). (a) Water temperature (b) Dissolved oxygen (c) Calcium hardness (d) Total dissolved solids.

Ecological variable selection for regression analyses Principal component Analysis (PCA) a multivariate statistic technique was used to convert a set of correlated environmental variables into a set of orthogonal, uncorrelated axes called principal components. The first three principal components accounted for 83.03% of the total variation in environmental covariates (Appendix S3 Figure 1). These components were related to stream characteristics i.e. canopy cover (PC1), water quality (PC2) and substrate type (PC3). Total alkalinity (TA), water

85 temperature (WT), total dissolved solids (TDS), dissolved oxygen (DO), electrical conductivity (EC) and pH were correlated strongly with PC2. Based on the PCA loadings I extracted variables such as water temperature, calcium hardness (CH), total dissolved solids, dissolved oxygen, canopy cover, depth-width ratio, and substrate type (% of pebble), to use as independent variables in exploring further regression analyses with guild-wise species richness as response variables (Table4. 2).

Table 4. 2. Comparison of environmental variable values in NRSB and RSB. Statistical significance (***) was obtained for differences in means using a Welch t- test (with unequal variance for groups) [alpha=0.05; p < 0.001***, ^ indicates no significant difference]. Environ variables Mean (SD) Welch t-test statistics

NRSB RSB df t p-value

Water temperature (0C) 20.86 24.76 45.2 -8.52 <0.001 ***

(2.28) (24.76)

TDS (mgL-1) 0.011 0.052 25.27 -6.15 <0.001 ***

(0.004) (0.033)

Calcium hardness (mgL-1) 8.69 110.69 21.16 -4.41 0.0002 ***

(7.93) (109)

DO (mgL-1) 7.80 8.35 27.77 -1.28 0.21^

(0.94) (1.64)

Inorganic nitrates (mgL-1) 0.17 0.38 23.09 -3.13 0.0046

(0.08) (0.29)

Differences in environmental variables in relation to river regulation and disturbance Water temperature was higher in regulated sub-basin than non-regulated sub-basin (df = 45.2, t =-8.52, p =0.001) and (Table 4. 2, Figure 4. 6) due to the lack of riparian vegetative cover in RSB than NRSB. Calcium hardness (df = 21.16, t =-4.41, p =0.002) and inorganic nitrates (df = 23.09, t =-3.13, p =0.004) were higher in RSB, indicating poorer water quality due to weathering of rocks (limestone, sedimentary rocks or calcium bearing minerals) and sewage generated from the agricultural and

86 town (Table 4. 2). Total Dissolved Solids (TDS) also followed a similar pattern (df = 25.27, t=-6.15, p <0.001), but dissolved oxygen (df = 27.77, t=-1.28, p =0.21) did not differ significantly between RSB and NRSB (Table 4. 2, Figure 4. 6). I did not detect any significant differences in water quality variables in relation to local water abstraction.

Effects of environmental variables on guild-wise fish species richness Responses to environmental variables differed across the three guilds (Table 4. 3). Overall, calcium hardness was negatively correlated and total dissolved solid concentration positively correlated species richness of all guilds. In NRSB, only water temperature positively influenced surface-dwelling fish species richness (Table 4. 3). In RSB, water temperature was positively correlated with species richness of BD and MCD guilds (Figure 4. 5, Table 4. 3). In summary, these results indicated that 1) river regulation influenced fish species richness differentially, by effecting changes in water temperature and water quality, and 2) water quality had a significant influence on fish richness in regulated sub-basins.

Discussion I found distinct differences in fish guild richness, species composition, and presence of endemic species across regulated and non-regulated sub-basins of Mhadei, Malaprabha, Tunga and Bhadra in the WG’s of India. Variable effects of flow regulation on different fish guilds were likely mediated by altered water temperature and chemical characteristics (water quality). Surface-dwelling fishes appeared to be affected by flow regulation indirectly through water quality, and their responses were different from mid-column and bottom-dwelling fishes. Water temperature was lower in NRSB compared to RSB, it is likely that SD fish richness increased till an upper limit, but declined at even higher temperatures (in RSB). For BD and MCD guilds, higher temperatures in RSB were perhaps more suitable. Surface-dwelling rheophilic and opportunistic fishes are known to require run-riffle habitats with good water quality in floodplain rivers for spawning (Costa et al., 2013; Borges and Araujo, 2013; Winemiller and Jasper, 1998). As a result, their lower richness in regulated sub-basins might be due to the combined effects of temperature and water quality. The amount of total dissolved solids (TDS), both in RSB and NRSB, was positively correlated with

Table 4. 3. Selected Generalized Linear Models (GLMs with Poisson errors) showing effect sizes of environmental variables on guild-wise species richness in all sub-basins, and for RSB and NRSB. (Sub-basins: RSB=regulated, NRSB=non-regulated; Variables: TDS=total dissolved solids, CH=Calcium hardness, WT=water temperature, DO=dissolved oxygen). Delta AIC values indicate improvement brought in fit by GLM over null (intercept-only) model. In case of MCD species richness in NRSB, no model was found to improve fit over the null model.

Guild Sub WT CH TDS DO Mac- AIC Delta AIC Sub-basin Estimates (SE) Estimates (SE) Estimates (SE) Estimates (SE) Fadden’s (w.r.t null Pseudo model) R2 -- SD All - -0.006 (-0.001) ** - - 0.24 199.63 217.2 RSB - - 40.94 (18.48) - 0.40 80.445 1.315 TDS2 -289.90 (139.76)- NRSB 0.06 (0.02)* - - - 0.09 321.57 10.51 MCD All - -0.02 (0.001). 12.03 (1.92) *** 0.15 (0.05)* 0.62 153.2 396.8 RSB 0.12 (0.04) ** -0.009 (0.001) *** - - 0.74 91.44 76.43 NRSB - - - - - null model BD All 0.081 (0.03) ** -0.003 (0.001) ** 10.18 (1.80) *** - 0.44 253.33 333.2 RSB 0.14 (0.03) *** -0.006 (0.001) *** - - 0.52 113.37 60.07 NRSB 0.05 (0.03). - - - 0.07 335.37 9.95 Statistical significance, alpha = 0.05: (p values * p <0.05, ** p <0.01, *** p <0.001).

BD and MCD guilds, but showed a quadratic relationship with surface-dwellers indicating that the SD guild was more sensitive to TDS than BD and MCD.

Water quality might decline after regulation due to changes in local land uses, human disturbance, or intensification of cropping patterns, as was observed in the study area (Heller et al., 2012). Poorer water quality could have also affected some endangered rheophilic fishes such as Tor sp, Barilius sp, and Hypselobarbus sp (field observations). Further, fishes with deformities were recorded only from RSB and never from NRSB (field observations). Lower water releases in RSB could have aggravated pollution impacts by not allowing adequate dilution or flushing of pollutants. NRSB river segments supported the persistence of many endemic and specialized fishes such as Balitora sp, Barbodes sp, Batasio sp, Rohtee sp etc. that were not detected in RSB. The consistently poorer water quality indicators in regulated sub-basins emphasized that monitoring of water quality as part of ecological flow regime maintenance which will be essential for the persistence of fish species impacted by river regulation.

My study could not detect any direct responses in terms of fish species richness to local water abstraction, for any of the guilds studied. Although certain specific effects are likely on fish abundance due to local changes in stream substrate, pollution sources, and minor disturbances for household uses of water. These effects could not be unraveled by this study for species richness. Future studies might thus need to look at effects of local disturbances on fish abundance and turnover than only on species richness.

I chose water column position-based guilds of fishes to examine their responses to flow regulation and associated environmental characteristics. This choice appeared to overcome potential redundancies in other guild classifications (e.g. fish communities dominated by omnivores), while still retaining correlations of fish position with other life-history strategies (particularly in reproductive guilds and sensitivity to flow alterations). Based on the exploratory analyses, I advocate the use of formal methods such as ordination for selecting appropriate guilds for analyzing fish species responses to regulation, based on available data.

Overall, the results matched with reported impacts of flow regulation on fish species. For instance, in the Teesta and Bhadra rivers in India, Connecticut stream in USA and few northern European large rivers, surface-dwelling rheophilic fishes were less abundant in regulated basins whereas eurytopic fishes in mid-column and bottom- dwelling guilds may have benefitted (David, 1956; Aarts and Nienhuis, 2003; De Leeuw et al., 2007; Chakrabarty and Homechaudhuri, 2013; Macnaughton et al., 2016). I observed that flow regulation modified connected river habitats to disconnected pools, which might have led to the above observations. The discrete fish guilds did not allow me to interpret a clear continuum in life-history strategies utilized by fishes as shown in previous studies from American rivers (Winemiller et al., 1989; 1992), but nonetheless helped detect broad differences. Studies on assessing the impact of dams on fish communities in rivers across South Asia have suggested that the dam causes considerable decline in native species by obstructing their seasonal spawning migrations (Jackson and Marmulla, 2001; Larinier, 2001; Hoeinghaus et al., 2009; Gopal, 2013). Upstream migration of surface-dwelling, rheophilic fishes such as Tor khudree, Hypselobarbus jerdoni and Cirrhinus fulungee could be affected due to hydrological barriers. Effects of existing barriers on changes in native fish breeding patterns also need to be studied in detail (Larinier, 2001).

River regulations due to barriers have disconnected river habitats by encouraging intensive gravel and sand mining and water abstraction in some segments especially during low water levels, this could have enhanced the water temperature. The relative contribution of surface water to ground-water to downstream flow is likely to decrease downstream of the barriers which may result in change in water quality such as increase in calcium hardness and temperature (Wurts and Robert, 1992; Hunse, 2007; Malkhede, 2003). The guild composition in such segments was dominated by bottom-dwelling fishes that appeared tolerant to river regulation. Land-use changes associated with the construction of an inter-basin water canal, road constructions, and check-dam constructions in the RSB might also have increased sediment deposition in headwater reaches, likely affecting potential habitat for endemic fishes. Some endemic species such as Bhavania australis were observed in the upper reaches of Mhadei sub-basin (pers. comm. Vijay Mohan Raj) but not in the upper reaches of Malaprabha due to loss of headstream habitat near the canal construction site. Siltation and reduction in dissolved oxygen is also known to affect some sensitive fish

90 species (Bhavania sp, Balitora sp), which reproduce in highly oxygenated benthic habitats (Ganasan and Hughes, 1998). Individual fishes with deformities and disease were recorded only in the Malaprabha river. I also observed that construction of check dams and barrages in the Malaprabha basin, over time, resulted in local fishing communities having to travel upstream for about 30-40 km daily to meet their subsistence requirements, suggesting wider impacts of declines in fish abundance, which need further detailed study.

Implication for conservation of fish diversity in the Western Ghats This study contributes to the understanding of broad ecological responses of fish guilds, based on their position in the river water column, to changes in water quality due to effects of river flow regulation (Winemiller, 1989). Importantly, the study identified that regulated river sub-basins had poorer quality of river water (e.g. increased level of calcium hardness and total dissolved solids, etc.) that led to negative effects on surface-dwelling fishes. Future threats such as inter-basin water transfers (an head-water link between the Malaprabha and Mhadei) are likely to therefore cause further reduction in water quality and seriously affect freshwater fish diversity and endemic species in particular (Lynch et al., 2011; Konar et al., 2013; Araujo et al, 2015; Sa-Oliveira et al., 2015). The WG region is the second highest in dam densities in India and more than 352 small to medium hydropower projects (< 25 MW) are under consideration in the Karnataka state alone which will threaten river biodiversity significantly (Dandekar, P., pers.comm). However the impacts of local disturbance on fish guild richness in this study appeared equivocal. These results can help generate specific hypotheses to be tested in future assessments fish responses (abundance) to flow alterations and local disturbance in tropical river systems. They also highlight the potential of identifying and maintaining ecological flow regimes to improve water quality not only for fish fauna but also for local human users of water.

The WG’s region shows a trend of rapid urbanization and human population pressures are likely to cause further river regulation in the near future (McDonald et al., 2011; Konar et al., 2013). I hypothesize that in future, the modified flow regulation and water pollution effects below the barriers are likely to change thermal regime and water quality affecting sensitive fish guilds in the tropical river systems. With these

91 imminent pressures in mind, catchment-scale conservation planning to protect endemic and endangered fish species needs to prioritize water quality improvement by emphasizing at-source treatment of pollutants and providing adequate flows downstream of barrages and dams, to reduce pollution effects. Freshwater conservation planning and policy in the current scenario of river regulation in the WG’s could benefit from this study both to mitigate impacts of existing river barriers and to assess ecological flow requirements for sensitive fish species.

References

Aarts B. G. W., and Nienhuis P. H. 2003. Fish zonations and guilds as the basis for assessment of ecological integrity of large rivers. Hydrobiologia. Pp. 157-178.

Abraham R. K, and Kelkar N. 2012. Do terrestrial protected areas conserve freshwater fish diversity? Results from the Western Ghats of India. Oryx 46: 544-553.

Ahmad S, Muralidharan M, Venkateshwarlu M, and Arunachalam M. 2013. Distribution pattern, endemism, threat status and conservation measures of fishes in the Tunga and Bhadra rivers of Western Ghats, India. Environmental Biology of Fishes 96: 1245-1256.

Anderson E. P., Freeman M. C., and Pringle C. M. 2006. Ecological consequences of hydropower development in Central America: impacts of small dams and water diversion on neotropical stream fish assemblages. River Research and Applications 22: 397-411.

Araújo C. C., Wang, and John Y. 2015. The dammed river dolphins of Brazil: impacts and conservation. Oryx 49: 17-24.

Arunachalam M. 2000. Assemblage structure of stream fishes in the Western Ghats (India). Hydrobiologia 430: 1-31.

ASIH. 2003. Guideline for use of fishes in field research. USA. 14.

Atkore, V., N. Kelkar and J. Krishnsawamy. 2017. Assessing the recovery of fish assemblages downstream of hydrological barriers in India's Western Ghats. River Research and Application. Doi:10.1002/rra.3163.

Baron J. S., N. LeRoy Poff, Paul L. Angermeier, Clifford N. Dahm, Peter H. Gleick, Nelson G. Hairston, Jr., Robert B. Jackson, Carol A. Johnston, Brian D. Richter, and Alan D. Steinman. 2003. Sustaining Healthy Freshwater Ecosystems. Issues in Ecology: 1-18.

Bhat A. 2002. A study of the diversity and ecology of freshwater ﬁshes of four river systems of the Uttara Kannada District, Karnataka, India. 178.

Bhat A. 2004. Patterns in the distribution of freshwater fishes in rivers of Central Western Ghats, India and their associations with environmental gradients. Hydrobiologia 529: 83-97.

Bhat A. 2005. Ecomorphological correlates in tropical stream fishes of southern India. Environmental Biology of Fishes 73: 211-225.

Bhat A, and Magurran A. E. 2007. Does disturbance affect the structure of tropical fish assemblages? A test using null models. Journal of Fish Biology 70: 623- 629.

Bhimachar B. S. 1945. Zoogeographical divisions of the Western Ghats as evidenced by the distribution of hill stream fishes. Current Science 1: 12-16.

Borcard D, Gillet F., and Legendre P. 2011. Numerical Ecology with R. Springer, Montréal, QC and Besançon, France. Pp.319.

Brasher A. M. D. 2003. Impacts of Human Disturbances on Biotic Communities in Hawaiian Streams. BioScience 53: 1052-1060.

Bunn E. S., and Arthington H. A. 2002. Basic Principles and Ecological Consequences of Altered Flow Regimes for Aquatic Biodiversity. Environmental Management 30: 492-507.

Chakrabarty M, and Homechaudhuri S. 2013. Fish guild structure along a longitudinally-determined ecological zonation of Teesta, an eastern Himalayan river in West Bengal, India. Arxius de Miscel-lania Zoologica 11: 196-213.

Chu L, Wang W, Zhu R, Yan Y, Chen Y, and Wang L. 2015. Variation in fish assemblages across impoundments of low-head dams in headwater streams of the Qingyi River, China: effects of abiotic factors and native invaders. Environmental Biology of Fishes 98: 101-112.

Cincotta R. P., Wisnewski J, and Engelman R. 2000. Human population in the biodiversity hotspots. Nature 404: 990-992.

Costa M.R. de, Mattos T. M., Borges J. L, and Araújo F. G. 2013. Habitat preferences of common native fishes in a tropical river in Southeastern Brazil. Neotropical Ichthyology 11: 871-880.

Dahanukar N, and Raghavan R. 2013. Freshwater fishes of the Western Ghats: Checklist v1.0 August 2013. Min 01: 6-16.

Daniels R. J. R. 2002. Freshwater fishes of Peninsular India. University Press India `(Pvt) Ltd, Hyderabad.Pp. 321.

Darwall W, Smith K, Allen D, Seddon M, Mc Gregor Reid, G. C, V., and Kalkman V. 2008. Freshwater biodiversity – a hidden resource under threat. In: J.-C. Vié, C. Hilton-Taylor and S.N. Stuart (eds.). The 2008 review of the IUCN Red

List of threatened species. Gland, Switzerland.

Das A, Krishnaswamy J, Bawa K. S., Kiran M. C, Srinivas V, Kumar N. S., and Karanth K. U. 2006. Prioritisation of conservation areas in the Western Ghats, India. Biological Conservation 133: 16-31.

David A. 1956. Studies on the pollution of the Bhadra river fisheries at Bhadravathi (Mysore state) with Industrial effluents. Proceedings of the National Academy of Sciences 22: 1-38.

De Leeuw J. J., Buijse A. D., Haidvogl G, Lapinska M, Noble R, Repecka R, Virbickas T, WiŚ Niewolski W, andWolter C. 2007. Challenges in developing fish-based ecological assessment methods for large floodplain rivers. Fisheries Management and Ecology 14: 483-494.

De Silva S. S., Cumaranatunga P. R. T., and De Silva C. D. 1979. Food, Feeding Ecology and Morphological Features Associated With Feeding of Four Co- Occurring Cyprinids (Pisces: Cyprinidae). Netherlands Journal of Zoology 30: 54-73.

Deacon A. E., Shimadzu H, Dornelas M, Ramnarine I. W., and Magurran A. E. 2015. From species to communities: the signature of recreational use on a tropical river ecosystem. Ecology and Evolution 5: 5561-5572. Dudgeon D. 2000. Large-Scale Hydrological Changes in Tropical Asia: Prospects for Riverine Biodiversity: The construction of large dams will have an impact on the biodiversity of tropical Asian rivers and their associated wetlands. BioScience 50: 793-806.

Dudgeon D. 2010. Prospects for sustaining freshwater biodiversity in the 21st century: linking ecosystem structure and function. Current Opinion in Environmental Sustainability 2: 422-430.

Dudgeon D, Arthington A. H., Gessner M. O., Kawabata Z-I, Knowler D. J., Lévêque C, Naiman R. J., Prieur-Richard A-H, eacute, egrave, ne, Soto D, Stiassny M. L. J., and Sullivan C. A. 2006. Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews 81: 163-182.

Eschmeyer W. N. 2015. Catalog of Fishes.in C. A. o. Sciences, editor. California Academy of Sciences.

Froese R, Pauly D. 2016. Fishbase. World Wide Web electronic publication.

Fuka D. R., Walter M. T., Archibald J. A., Steenhuis T. S., and Z. M. Easton. 2014. A community modeling foundation for Eco-Hydrology.

Ganasan V., and Hughes R. M. 1998. Application of an index of biological integrity (IBI) to fish assemblages of the rivers Khan and Kshipra (Madhya Pradesh), India. Freshwater Biology 40: 367-383.

Gandini C. V., Sampaio F. A. C., and Pompeu P. S. 2014. Hydropeaking effects of on

the diet of a Neotropical fish community. Neotropical Ichthyology 12: 795- 802.

Gilliam J. F., Fraser D. F., and Alkins-Koo M. 1993. Structure of a Tropical Stream Fish Community: A Role for Biotic Interactions. Ecology 74: 1856-1870.

Gopal B. 2013. Environmental Flow: An introduction for Water Resource Managers. National Institute of Ecology (URL: www.nieindia.org), 366 Metro Apartments, DDA Flats, Jahangirpuri, Delhi 110033, India. Email: [email protected].

Heller E, J. M. Rhemtulla, S. Lele, M. Kalacska, S. Badiger, and RS and Ramankutty N. 2012. Mapping crop types, irrigated areas and cropping intensities in heterogenous landscapes of southern India using multi-temporal medium resolution imagery: Implications for assessing water use in agriculture. Photogrammetric engineering and Remote Sensing 78: 815-827.

Hoeinghaus D. J., Agostinho A. A., Gomes L. C., Pelicice F. M., Okada E. K., Latini J. D., Kashiwaqui E. A. L., and Winemiller KO. 2009. Effects of River Impoundment on Ecosystem Services of Large Tropical Rivers: Embodied Energy and Market Value of Artisanal Fisheries. Conservation Biology 23: 1222-1231.

Hunse T. H. 2007. Ground water information booklet. Belgaum district, Karnataka. CentralGround Water Board, Ministry of Water Resources, Government of India., South Western Region Bangalore. 30.

Ibrahampurkar M. M. 2012. Hydrological and hydrogeological evaluation of Mhadei river watershed in Goa and Karnataka. A Ph.D thesis submitted to The Department of Earth Science. Goa University, Panjim. 282.

IUCN. 2011. Guidelines for appropriate uses of IUCN Red List Data. Incorporating the Guidelines for Reporting on Proportion Threatened and the Guidelines on Scientific Collecting of Threatened Species. Version 2. Adopted by the IUCN Red List Committee and IUCN SSC Steering Committee. Downloadable from: http://intranet.iucn.org/webfiles/doc/SpeciesProg/RL_Guidelines_Data_Use.p df or http://www.iucnredlist.org/documents/RL_Guidelines_Data_Use.pdf.

Jackson D. A., Peres-Neto P. R., and Olden J. D. 2001. What controls who is where in freshwater fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries and Aquatic Sciences 58: 157-170.

Jackson D. C., and Marmulla G. 2001. The influence of dams on river fisheries. In Dams, fish and fisheries. Opportunities, challenges and conflict resolution. Ed. by G. Marmulla. FAO Fisheries Technical Paper. No. 419. Rome, FAO. 2001. 166p.: 1-44.

Jayaram K. C. 2010. The Frehswater Fishes of Indian Region. Narendra Publishing House,Delhi.Pp. 551.

Johnson J. A., and Arunachalam M. 2012. Feeding habit and food partitioning in a stream fish community of Western Ghats, India. Environmental Biology of Fishes 93: 51-60.

Konar M, Jason Todd M, Muneepeerakul R, Rinaldo A, and Rodriguez-Iturbe I. 2013. Hydrology as a driver of biodiversity: Controls on carrying capacity, niche formation, and dispersal. Advances in Water Resources 51: 317-325.

Krishnaswamy J, M. Bunyan, V. K. Mehta, and NJ, Karanth KU. 2006. Impact of iron ore mining on suspended sediment responses in a tropical catchment in Kudremukh, Western Ghats, India. Forest Ecology and Management 224: 187-198.

Kundu N, D. Mazumdar,Homechaudhuri S. 2014. Seasonal changes in fish assemblages and trophic guild in selected interconnected rivers in eastern India. Indian journal of Fisheries 61: 35-44.

Larinier M. 2001. Environmental issues, dams and fish migration. In Dams, fish and fisheries. Opportunities, challenges and conflict resolution. Ed. by G. Marmulla. FAO Fisheries Technical Paper. No. 419. Rome, FAO. 2001. 166p.: 45-90.

Lowe-McConnell R. H. 1975. Fish communities in tropical freshwaters their distribution, ecology and evolution. London.

Lowe-McConnell R. H. 1987. Ecological studies in tropical fish communities. Cambridge University Press, Cambridge, Cambridge.

Lynch H. J., Campbell Grant E. H., Muneepeerakul R, Arunachalam M, Rodriguez- Iturbe I,and Fagan W. F. 2011. How restructuring river connectivity changes freshwater fish biodiversity and biogeography. Water Resources Research 47: n/a-n/a.

Macnaughton C. J., Senay C, Dolinsek I, Bourque G, Maheu A, Lanthier G, Harvey- Lavoie S, Asselin J, Legendre P, and Boisclair D. 2016. Using fish guilds to assess community responses to temperature and flow regimes in unregulated and regulated Canadian rivers. Freshwater Biology 61: 1759-1772.

Malkhede SK. 2003. Working Plan for Belgaum Forest Division (2002-2003 to 2011- 2012). Belgaum Forest Division, Belgaum, Belgaum, Karnataka state, India. 360.

Malmqvist B, Rundle S. 2002. Threats to the running water ecosystems of the world. Environmental Conservation 29: 134-153.

McDonald R. I., Green P, Balk D, Fekete B. M., Revenga C, Todd M, and Montgomery M. 2011. Urban growth, climate change, and freshwater availability. Proceedings of the National Academy of Sciences 108: 6312- 6317.

Merz S. K. 2013. Characterizing the relationship between water quality and water quantity. St Leonards NSW 2065 Australia. 139.

Mims M. C., and Olden J. D. 2012. Life history theory predicts fish assemblage response to hydrologic regimes. Ecology 93: 35-45.

Mims M. C.,and Olden J. D. 2013. Fish assemblages respond to altered flow regimes via ecological filtering of life history strategies. Freshwater Biology 58: 50- 62.

Moreno C. E., Rodriguez P. 2010. A consistent terminology for quantifying species diverstity? Oecologia 163: 279-282.

Mouillot D, Graham N. A. J., Villéger S, Mason N. W. H., and Bellwood D. R. 2013. A functional approach reveals community responses to disturbances. Pages 167-177 Trends in Ecology and Evolution. Elsevier.

Nel J. L., Roux D. J., Maree G, Kleynhans C. J., Moolman J, Reyers B, Rouget M, and Cowling R. M. 2007. Rivers in peril inside and outside protected areas: a systematic approach to conservation assessment of river ecosystems. Diversity and Distributions 13: 341-352.

Oliveira J. M., Segurado P, Santos J. M., Teixeira A, Ferreira M. T.,and Cortes R. V. 2012.Modelling Stream-Fish Functional Traits in Reference Conditions: Regional and Local Environmental Correlates. PLoS ONE 7: e45787.

Poff N. L., Allan J. D. 1995. Functional Organization of Stream Fish Assemblages in Relation to Hydrological Variability. Ecology 76: 606-627.

Poff N. L., and Zimmerman J. K. H. 2010. Ecological responses to altered flow regimes: a iterature review to inform the science and management of environmental flows. Freshwater Biology 55: 194-205.

Pohlert T. 2016. Trend: Non-Parametric Trend Tests and Change-Point Detection.

Pool T. K., Olden J. D., Whittier J. B., and Paukert C. P. 2010. Environmental drivers of fish functional diversity and composition in the Lower Colorado River Basin. Canadian Journal of Fisheries and Aquatic Sciences 67: 1791-1807.

Pringle C. M. 2001. Hydrologic connectivity and the management of biological reserves: a global perspective. Ecological Application. 11: 981-998.

Raghavan R, Das S, Nameer PO, Bijukumar A, and Dahanukar N. 2016. Protected areas and imperilled endemic freshwater biodiversity in the Western Ghats Hotspot. Aquatic Conservation: Marine and Freshwater Ecosystems 26: 78- 90.

Raghavan R, Prasad G, Ali P. H. A., and Pereira B. 2008. Fish fauna of Chalakudy River, part of Western Ghats biodiversity hotspot, Kerala, India: patterns of distribution, threats and conservation needs. Biodiversity and Conservation

17: 3119-3131.

Reid GM, Contreras MacBeath T, and Csatádi K. 2013. Global challenges in freshwater-fish conservation related to public aquariums and the aquarium industry. International Zoo Yearbook 47: 6-45.

Rema Devi K, Indra T. J., Yadav B. E., Raghunathan M.B., Krishnan S., and Jadhav S. S. 2013. Freshwater fishes In 'State Fauna Series. Fauna of Karnataka' Pages 1-595 in Z. S. I. Editor-Director, editor. State Fauna Series. Zoological Survey of India, Kolkata.

Rowe D. C. 2007. Relationships of fish assemblages, instream physical habitat, and landscape characteristics of wadeable Iowa streams. Retrospective Theses and Dissertations. Paper 14819. Iowa State University. 145.

Sa-Oliveira Julio C., Issac V. J., and Ferrari S. F. 2015. Fish community structure as an indicator of the long-term effects of the damming of an Amazonian river. Environmental Biology of Fishes 98: 273-286.

Snyder D. E. 2003. Electrofishing and its harmful effects on fish. Information and Technology Report 2003-0002. U.S. Geological Survey. 149.

Strayer D. L. and D. Dudgeon. 2010. Freshwater biodiversity conservation:recent progress and future challenges. Journal of North American Benthological Society 29: 344-358.

Team R Development Core. 2013. A language and environment for statistical computing., R Foundation for Statistical Computing, Vienna, Austria.

Vannote R. L. G. W. M., K.W. Cummins, J. R. Sedell, and C. E. Cushing. 1980. The River Continuum Concept. Canadian Journal of Fisheries and Aquatic Sciences: 130-137.

Vasconcelos L. P., Alves D. C., and Gomes L. C. 2014. Fish reproductive guilds downstream of dams. Journal of Fish Biology 85: 1489-1506.

Villéger S. 2008. Dynamique de la diversité fonctionnelle des communautés de poissons (Lagune de Terminos, Mexique). Montpellier SupAgro. 328p.

Vorosmarty C. J., McIntyre P. B., Gessner M. O., Dudgeon D, Prusevich A, Green P, Glidden S, Bunn S. E., Sullivan C. A., Liermann C.R., and Davies PM. 2010. Global threats to human water security and river biodiversity. Nature 467: 555-561.

Welcomme R. L. 1985. River Fisheries. FAO. 330.

Welcomme R. L., Winemiller K. O., and Cowx IG. 2006. Fish environmental guilds as a tool for assessment of ecological condition of rivers. River Research and Applications 22: 377-396.

Weliange W. S., and Amarsinghe U. S. 2007. Relationship between body shape and foodhabits of fish from three reservoirs of Sri Lanka. Asian Fisheries Science 20: 257-270.

Winemiller K. O. 1989. Patterns of variation in life history among South American fishes in seasonal environments. Oecologia 81: 225-241.

Winemiller K. O., and Jepsen D. B. 1998. Effects of seasonality and fish movement on tropical river food webs. Journal of Fish Biology 53: 267-296.

Winemiller K. O, and Rose K. A. 1992. Patterns of Life-History Diversification in North American Fishes: implications for Population Regulation. Canadian Journal of Fisheries and Aquatic Sciences 49: 2196-2218.

Winemiller K. O., Agostinho A. A., and Pellegrini E. C. 2008. Fish ecology in tropical streams. in D. Dudgeon, editor. Tropical Stream Ecology. Elsevier/Academic Press, San Diego, CA. Pages 107-146.

Wurts W. A., and Robert M. D. 1992. Interactions of pH, carbon dioxide, alkalinity and hardness in fish ponds. Southern Regional Aquaculture Center, USA: 1-4.

Yatagai A, Kamiguchi K, Arakawa O, Hamada A, Yasutomi N, and Kitoh A. 2012. APHRODITE: Constructing a Long-Term Daily Gridded Precipitation Dataset for Asia Based on a Dense Network of Rain Gauges. Bulletin of the American Meteorological Society 93: 1401-1415.

Yue S, Pilon P, and Cavadias G. 2002. Power of the Mann–Kendall and Spearman's rho tests for detecting monotonic trends in hydrological series. Journal of Hydrology 259: 254-271.

Chapter 5

Assessing the recovery of fish assemblages downstream of hydrological barriers in India’s Western Ghats

Publication Atkore, V., Nachiket Kelkar & Jagdish Krishnaswamy 2017. Assessing the recovery of fish assemblages downstream of hydrological barriers in India’s Western Ghats. River Research & Application. Doi:10.1002/rra.3163.

100

Assessing the recovery of fish assemblages downstream of hydrological barriers in India’s Western Ghats

Introduction Anthropogenic flow regulation and the resulting fragmentation of stream and river habitats is a major threat to conservation of freshwater fish species worldwide (Bunn and Arthington, 2002; Mamqvist and Rundle, 2002; Nilsson et al., 2005; Dudgeon, 2010; Vorosmarty et al., 2010). Downstream impacts of ‘river barriers’ such as dams, barrages, and so-called ‘run-of-river’ hydropower projects are often severe for tropical fish communities (Pringle, 2001; Pandit and Grumbine, 2012; Anderson et al., 2006, Brown et al., 2013). Presence of barriers might thus affect fish recovery at different spatial extents and scales within their catchments (Dudgeon, 2011). Flow alterations not only reduce stream discharge needed to maintain fish abundance and diversity downstream of barriers, but can also induce fluctuations and disturbances, to natural flow regimes downstream, causing habitat loss, mass mortality, and eventual local extinction (Travnichek et al., 1995; Bunn and Arthington, 2002; Phomikong et al., 2015). Despite the well-documented threats from these interventions (Pringle, 2001; Hoeinghaus et al., 2009; Dudgeon, 2010; Sakaris, 2013), the potential for ecological recovery of fish communities below barriers, and the biotic and abiotic factors contributing to recovery, are not well understood (Ellis and Jones, 2013), in part due to lack of empirical data, especially for species rich tropical rivers.

Regulated rivers are distinctly different in their ecological processes when compared to natural, free-flowing rivers. Theoretically, this difference has been attributed to hydrological connectivity in free-flowing rivers - the ‘River Continuum’ concept; Vannote et al (1980) by which natural gradients in species composition and community structure of aquatic organisms are maintained. In contrast, in regulated rivers, habitat diversity resulting from discontinuities in the river flow determines spatial turnover in fish diversity, as understood by the ‘Serial Discontinuity’ concept (Ward and Stanford, 1995). The concept proves especially useful in understanding processes of recovery following disruptions in longitudinal and lateral connectivity caused by human-made river barriers (Ellis and Jones, 2013). Further work attributes

101 a greater role to functional processes (Thorp et al., 2008) in locally unaltered habitat nodes within the larger modified catchment or stream network, especially as the distance from the barriers increases and the impact of flow impairment reduces. Beyond certain distance thresholds, basin-level hydrological connectivity might improve, and thus allow some stretches to return to near-natural baseline conditions where species might persist (Connell and Sousa, 1983). In addition, some species’ life-history traits may be more resilient to flow alterations than others, and their potential for recovery could be higher than sensitive species.

Central to the identification of factors enabling recovery is understanding the complex and nonlinear relationships between river flow regimes, habitat quality, and responses of fish biodiversity (Taylor et al., 2014; King et al., 2015; Macnaughton et al., 2015). Kubach, Scott and Bulak (2011) reported recovery in fish assemblages after a 52- month period following an oil spill in the Reedy River in USA, indicating that recovery might be observed after long time-lags, rather than being determined by only spatial location effects. The persistence of fluvial specialists (rheophilic fishes with high swimming ability) downstream of dams is generally considered as an important part of recovery (Travnichek et al., 1995). Increase in depth and discharge downstream is also known to contribute to recovery of rheophilic species, the strength of recovery typically increasing with downstream distance from the barrier (Scheidegger and Bain, 1995; Stoll et al., 2014; Piller and Geheber, 2015). Though overall fish species richness increases downstream (Ibañez et al., 2009), this is purely due to spatial turnover: rheophilic species are replaced by eurytopic species that are well adapted to lentic habitats formed by flow reduction (Scheidegger and Bain, 1995; Aarts and Nienhuis, 2003). Recovery might also be affected by topographic factors, land-use in the catchment, or other human disturbances such as pollution and water abstraction (Jackson et al., 2001a; Schlosser, 1991). As a result, separating the effect of species replacement (turnover) from recovery, while examining patterns in species community composition, is important (Kubach et al., 2011). More recent studies have focused on the contribution of undammed downstream tributaries to functional connectivity and ability of fish species with different life-history traits to recover below large barriers (Brown and Ford, 2002; Alexandre et al., 2013). Fish swimming ability and movement are key determinants of recovery, as they directly influence how and whether fish can negotiate barriers when dam/barrage gates may be opened

102

(e.g. during seasonal flooding) to travel upstream or downstream (Rolls and Sternberg, 2015).

Studies on fish recovery are relatively few from tropical South Asia, despite the high density of and intensive river flow regulation by existing and under-construction dams, barrages, and hydropower projects on rivers (Bhat and Magurran, 2007; Pandit and Grumbine, 2012; Theophilus, 2014). In this chapter, I study the ecological processes contributing to fish species recovery below large barrages (height > 2 m, width 50-150 m, and > 10 km2 impoundment area) and small barriers (small barrages and check-dams =< 2 m, < 50 m width and < 1.5 km2 impoundment area) in the upper catchment of the Malaprabha River of the WG’s of India. There have been recent inter-state conflicts over water sharing through inter-basin transfers in this river basin, and demands are intensifying. As a result, conservation planning for freshwater taxa needs to be prioritized in this eco-region. In this regard, studies on fish responses to flow regulation can help assess the cumulative impacts of current and future water developments on regulated stream networks.

Here, I assess the potential for recovery of native fish assemblages downstream of river barriers. For this I define ‘recovery’ as the similarity in fish community composition as compared to the composition in undammed reaches upstream of all barriers within the basin. I hypothesize that the presence of undammed streams joining rivers downstream of barriers will have an amelioration effect and positively influence species recovery. Based on this hypothesis, I ask two broad questions: (1) what is the potential for species recovery below barriers? (2) what river habitat characteristics are likely to influence spatial patterns in recovery (correlated with downstream distance from barriers)? As per the Serial Discontinuity Concept (Stanford and Ward, 2001), I predict that similarity in species composition will 1) reduce initially, and 2) increase subsequently with increasing distance from the barrier, owing to undammed tributaries joining downstream (Fig. 5.1a, b). Finally, I discuss implications of fish recovery for aquatic conservation planning, restoration of river habitats, and planning ecological flow regimes in regulated river basins of the WG.

103

Figure 5. 1. (a) Schematic showing the application of the Serial Discontinuity Concept (SDC) in arriving at the hypotheses tested in our study. (b) Expected pattern of change in species recovery with increasing distance downstream of barriers.

Materials and Methods Study Site The Central and Southern WGs of India have high freshwater fish diversity and endemism (Bhat, 2004; Raghavan et al., 2008; Molur et al., 2011). The east-flowing Malaprabha river originates from Kankumbi in the central WG in the state of Karnataka, India (Fig. 5. 2). Annual rainfall recorded in the Malaprabha ranged from 2000 to 3500 mm and mean discharge from 1000 to 2300 m/s (Jha & Singh, 2008). An overall negative trend in discharge was noted post-regulation, despite no appreciable reduction in rainfall, suggesting the potential for regulation impacts on fish guild structure (V. Atkore, unpublished data). The main land-use types in sub- basin are forests and agriculture with sugarcane, sunflower, and paddy as the main crops. Water abstraction by pumps, illegal sand/boulder mining, pollution, and destructive fishing methods are also intensifying threats in the basin. The river is highly regulated by barrages and check-dams, used both for irrigation and domestic purposes. I identified 10 (1 large, 9 small) barriers in the upper catchment of the basin (> 500 m ASL elevation), (Table 5. 1). An inter-basin link between the Malaprabha and Mhadei river (a west flowing river) has affected some of the crucial riparian habitats for stream fishes in the headwater reaches (Atkore et al., 2012).

104

Table 5. 1. Site-wise details on locations of barriers, cumulative impacts and number of undammed tributaries present between segments and their nearest upstream barriers.

River Coordinates No. of Distance from Cumul. No. of. Segment Lat. Long. barriers nearest upstream impact undam. (small, large) barrier (km) tributar.

Habbanhatti 15.71N, 74.35E 1, 0 0.29 3.45 0 Kusmali 15.71N, 74.37E 1, 0 4.70 0.2 1 Olmani 15.71N, 74.40E 1, 0 8.82 0.07 2 Shankerpeti 15.68N, 74.42E 4, 0 4.40 0.16 3 Malavi 15.65N, 74.43E 4, 0 0.12 25 0 Asoga 15.62N, 74.47E 5, 0 0.22 8.82 0 Shedegali 15.62N, 74.49E 6, 0 2.91 1.23 0 Rumewadi 15.63N, 74.51E 6, 1 0.53 0.78 0 Kupatgiri 15.63N, 74.53E 6, 1 2.30 0.57 1 Katgali 15.71N, 74.47E 1, 0 0.05 20 0 Valmiki 15.71N, 74.45E 1, 0 2.06 0.47 1

Figure 5. 2. Map showing fish species recovery (shown for total fish species) in the Malaprabha basin. Locations of upstream control segments (unregulated reference sites) and segments downstream of barrier (test segments) are shown in the inset. Sampling strategy

105

In this study I defined ‘river regulation’ mainly with reference to ‘large barriers’ (i.e. barrages), and ‘small barriers’ (i.e. temporary impoundments, small barrages, check dams) in the basin. I assessed patterns of spatial recovery and identified potential factors influencing recovery for total fish species richness and endemic fish species richness. I defined species recovery as being strictly downstream of these barriers, which I compared against a ‘reference composition’ derived from occurrence data only for species that were recorded in upstream control sites (river segments) that were not influenced by any barriers. A river segment (150 m long) was the sampling unit and a total of 15 segments were sampled, including 4 upstream control segments and 11 segments downstream of large and small barriers (Table 5. 1). I calculated stream order of the segments from a GIS map made by digitizing streams from Survey of India topographical maps with scale 1: 50,000. I also controlled for effects of elevation on species turnover by choosing all sites in the same elevation range. Both upstream and downstream segments mainly included river pool habitats, and had similar elevational extents (mean 693 ± SD 26 m) and stream orders (2-4). As background variables were similar, the upstream-downstream comparisons were justified and allowed for testing the specific impacts of river regulation by barriers. I restricted sampling to pools because these were the only habitat type consistently available for sampling, due to effects of dewatering and flow regulation.

Between the years 2011 and 2014, I sampled fishes using 0.5 cm x 0.5 cm and 1.1 cm x 1.1 cm mesh sized cast nets, using replicate casts to calculate species accumulation. Cast nets have been widely used for other studies in the WG and are known to be suitable for sampling different fish guilds in hill-stream sections (Bhat, 2004; Abraham and Kelkar, 2012). In addition, two visual transects were conducted along each segment to detect easily visible species on the surface and mid-column. Cast netting effort combined with visual records was deemed adequate for sampling all major fish guilds, i.e. feeding, positional, reproductive and flow dependent guilds (Welcomme et al., 2006) regardless of relative abundance (V. Atkore, unpublished data). About 80% of the species reported from the upper catchments of the Krishna river Basin (Rema Devi et al., 2013) were sampled in my study, indicating adequate sampling effort.

Environmental variables describing stream habitat and water quality characteristics

106 were measured for each segment. I recorded depth and channel width (m) of the stream manually using tape measures. I visually classified stream substrate and estimated percentages of different sediment categories (rocks, boulders, pebbles, cobbles, sand, and clay) for each segment at 3 locations along the stream length and averaged the percentage of each category later. Temperature was measured using a handheld thermometer. I measured different water quality variables in the field as follows, using standard estimation procedures as described in (APHA, 2005). 1) Dissolved Oxygen (DO) was measured using Winkler’s method 2) pH and electrical conductivity were measured using hand-held meters (HI98130 and HI98303), and 3) total alkalinity, calcium hardness, concentration of inorganic nitrates, phosphates, and chlorides were measured using laboratory analyses as prescribed by reference manuals of APHA (2005). Further, I derived an index of human disturbance level based on observed activities along the sampled segment (e.g. fishing, sand/boulder mining, domestic uses, pollution etc.). I used field surveys and GIS maps to spatially record the presence of adjacent undammed streams (tributaries) for each segment. I measured the exact downstream distance of the segment from the nearest large or small barriers in a GIS system. The ‘cumulative impact’ of these upstream barriers was scaled as the number of barriers per km of downstream distance from the nearest barrier (Table 5. 1). These variables were used as ecological predictors to test our hypotheses about fish species recovery (response variable).

Statistical Analysis For analysis, I first calculated the fish community similarity, i.e. the proportion of species occurring downstream of barriers, out of the total species and endemic species richness recorded in upstream control sites. I classified recovery into three levels: 1) low (< 0.33), 2) moderate (0.34-0.66) and 3) high (0.67-1.00). I used boxplots and scatter-plots to explore how species recovery changed with downstream distance of segments from barriers. I then used non-linear regressions to model the pattern of species recovery as a function of increasing downstream distance from barriers. I compared two types of non-linear models that reflected different predictions about how recovery was expected to increase: 1) recovery immediately rises to the highest possible after a threshold distance (asymptotic exponential model), and 2) recovery continues to increase at a different rate after the threshold distance (logistic model).

107

Asymptotic exponential models were of the form 푦 = 푎 × (1 − 푒−푏∗퐷푖푠푡푎푛푐푒), and

(푏−퐷푖푠푡푎푛푐푒) logistic models had the form y = a/(1 + 푒 푐 ) where y = total or endemic species recovery, and Distance = downstream distance from the barrier, and a, b, c were model parameters governing the shape of the function. No transformation of variables was required for parameter estimation. Model comparisons were based on how well the predicted and observed values corresponded (visual estimation of fit), and a statistical measure of fit (residual standard error of models) (Crawley, 2007). The Akaike Information Criterion (AIC) was used to guide model selection based on fit and parsimony, and simpler models were selected. All analyses were conducted in the R software package ‘nlme’ (R Core Team 2013). Recovery was assessed at two distance thresholds, one based on the median distance (2 km) for segments downstream from barriers, and at a greater arbitrarily determined threshold of 5 km. I qualitatively assessed if the presence of undammed tributaries joining the impaired main stem of the river improved fish recovery. I then ran Spearman’s rank correlation tests to assess how stream characteristics (water quality, stream morphology, etc.) correlated with downstream distance from barriers, and fish recovery. Life-history traits and habitat preferences of fish species were compiled from literature sources, to discuss what traits made them more or less able to recover post river regulation (Daniels, 2002; Albanese et al., 2009).

Results I sampled 28 species in the upstream control sites (n = 4) of the Malaprabha sub- basin, of which 14 were WG endemics (details in Table 5. 2). Bottom dwelling species were dominant (46.42%) followed by mid-column (32.14%) and surface dwelling species (21.42%) (Table 5. 2). The fish species monitored in the study were mainly rheophilic, periodically breeding, rock- or plant-spawners (Table 5. 2). Observed spatial patterns of fish community similarity (recovery) confirmed the hypothesis that recovery was low immediately downstream of barriers, and increased with increasing downstream distance (Fig. 5. 1, 5. 2). These patterns were consistent for endemic fish richness and for total fish species richness sampled in the upstream control sites (Fig. 5. 3a, 3b).

Recovery was negatively correlated with the number of upstream large and small

108 barriers (scaled to barriers per km) in the sub-basin, indicating their cumulative impacts (Fig. 5. 3c, Table 5. 1). Total species recovery was positively correlated with dissolved oxygen (Spearman’s rho = 0.84, p = 0.004) and rocky stream substrate (Spearman’s rho = 0.65, p = 0.028). Similarly, endemic species recovery was positively correlated with dissolved oxygen (Spearman’s rho = 0.93, p = 0.0002), water temperature (Spearman’s rho = 0.57, p =0.64), and rocky substrate (Spearman’s rho = 0.71, p = 0.014). Total (Spearman’s rho = -0.65, p = 0.03), and endemic species recovery (Spearman’s rho = -0.59, p = 0.05) were negatively correlated with total alkalinity. Of these, dissolved oxygen and total alkalinity respectively increased and decreased with distance from barriers (Fig. 5.3e, 3f). Water temperature and rocky habitat showed no correlation with distance from barriers (temperature: Spearman’s rho = 0.03, p = 0.92; rocky habitat: Spearman’s rho = 0.27, p = 0.40). The number of undammed tributaries (of stream order >= 3) downstream of barriers positively influenced recovery of species, as expected (Fig. 5. 3d). Both total and endemic species recovery increased to the highest level after a distance of 2 km downstream of barriers, as per the predictions of the non-linear asymptotic exponential models. Parameter estimates, model fit and selection criteria are given in Table 5. 3.

Out of 28 species, 26 recovered downstream of barriers, and two species did not. Till 2 km downstream, 24 species (85.7%) reappeared, of which 13 endemics out of 14 were recorded. Till the 5 km threshold, 25 out of 28 fish species (89.3%) with 13 endemics, were recorded. High levels of recovery were observed for most fish species that we monitored in the study, but no clear associations were noted with any particular guilds. Fishes that showed low recovery were typically with low swimming ability and benthic habitat preferences. These species included the loaches Nemacheilus thermalis and Paracanthocobitis mooreh, the catfish Clarias batrachus, and carps Hypselobarbus dobsoni and Chela cachius.

109

Figure 5. 3. Increase in (a) total and (b) endemic fish species recovery with increasing distance from upstream barriers; reduction in recovery (c) at higher levels of cumulative impact (number of upstream barriers/ downstream distance in km) and improvement in species recovery with the number of undammed tributaries joining the river below barriers (d). Dissolved oxygen (DO) (e) increased with distance from barrier, and total alkalinity (f) reduced with distance from barrier.

110

Table 5. 2. Species sampled in the study area, occurrence up to 2 and 5 km distances downstream of barriers, with life-history traits and recovery levels.

Fish species Distance Guild classifications (Order, Family) thresholds 2 5 Position in Flow Diet Swimming Life-history Reproduction- Recovery km km water column preferences ability strategy based Level Cypriniformes Cobitidae Paracanthocobitis mooreh* 1 0 Bottom Rheophilic Heterotrophic Low Periodic Lithophils Low Nemacheilus thermalis* 1 1 Bottom Rheophilic Phytophagous Low Periodic Lithophils High Cyprinidae Cirrhinus fulungee 1 1 Surface Rheophilic Omnivore High Periodic Lithophils High Chela cachius 0 0 Surface Rheophilic Omnivore Moderate Periodic Phytolithophils Low Devario malabaricus* 1 1 Surface Rheophilic Heterotrophic High Opportunistic Phytolithophils High Garra bicornuta* 1 1 Bottom Rheophilic Phytophagous Moderate Opportunistic Lithopelagophils High Garra mullya* 1 1 Bottom Rheophilic Phytophagous Moderate Opportunistic Lithopelagophils High Garra stenorhynchus* 1 1 Bottom Rheophilic Phytophagous Moderate Opportunistic Lithopelagophils High Hypselobarbus curmuca 1 1 Bottom Rheophilic Omnivore High Periodic Phytophils High Hypselobarbus dobsoni* 0 1 Bottom Rheophilic Phytophagous High Periodic Lithophils Moderate Osteochilius nashii* 1 1 Mid-column Rheophilic Omnivore High Periodic Lithophils High Pethia setnai* 1 1 Mid-column Rheophilic Heterotrophic Moderate Intermediate Phytophils High Pethia ticto 0 1 Mid-column Eurytopic Omnivore Moderate Periodic Phytophils Moderate Puntius amphibius* 1 1 Mid-column Rheophilic Heterotrophic Moderate Intermediate Phytophils High Puntius sophore 1 1 Mid-column Eurytopic Heterotrophic Moderate Periodic Phytophils High Rasbora daniconius 1 1 Mid-column Eurytopic Heterotrophic High Opportunistic Phytophils High Rasbora labiosa* 1 1 Mid-column Eurytopic Omnivore Moderate Opportunistic Phytophils High Salmophasia bacaila 1 1 Surface Limnophilic Omnivore High Periodic Phytophils High Salmophasia boopis 1 1 Surface Limnophilic Heterotrophic High Periodic Phytophils High Salmophasia novacula* 1 1 Surface Limnophilic Heterotrophic Moderate Periodic Phytophils High

111

Fish species Distance Guild classifications (Order, Family) thresholds 2 5 Position in Flow Diet Swimming Life-history Reproduction- Recovery km km water column preferences ability strategy based Level Systomus sarana 1 1 Mid-column Limnophilic Phytophagous High Intermediate Phytophils High Tor khudree* 1 1 Mid-column Rheophilic Omnivore High Periodic Lithophils High Siluriformes Bagridae Mystus bleekeri 1 1 Bottom Eurytopic Heterotrophic Moderate Equilibrium Phytophils High Mystus cavasius 1 1 Bottom Rheophilic Heterotrophic Moderate Periodic Phytophils High Mystus gulio 1 1 Bottom Rheophilic Heterotrophic Moderate Periodic Phytophils High Mystus keletius* 1 1 Bottom Rheophilic Heterotrophic Low Periodic Phytophils Low Clariidae Clarias batrachus 0 0 Bottom Eurytopic Omnivore Moderate Opportunistic Phytophils Low Siluridae Ompok bimaculatus 1 1 Bottom Eurytopic Heterotrophic High Periodic Polyphils High

Key: * Fish species endemic to the Western Ghats (Dahanukar & Raghavan, 2013; Daniels, 2002). Guild classifications are based on Daniels 2002, Welcomme et al., (2006), and Albanese et al., (2009). Flow preferences: Eurytopic=all life stages occur both in lotic and lentic waters, Limnophilic=all life stages confined to lentic water with macrophytes, Rheophilic=all life stages are confined to river flows in main channel. Diet: Heterotrophic=carnivores and insectivores, Phytophagous=algal, plant, and detritus feeders, Omnivores=both plant and animal matter is fed on. Life-history strategies: Equilibrium=fishes with high parental care and juvenile survival with intermediate maturation periods, Opportunistic=early maturing fishes with continuous reproduction but low fecundity, Periodic=large fishes with late maturation, high fecundity, and low juvenile survival, Intermediate=small fishes with prolonged and distinct seasonal reproduction. Reproduction-based guilds: Lithophils=rock/gravel spawners, Lithopelagophils=rock, gravel spawners with pelagic larvae, Phytolithophils=non-obligatory plant spawners, phytophils=obligatory plant spawners, Pelagophils=pelagic spawners, polyphils=no substrate preferences, and psammophils=sand spawners. Recovery level: Observed occurrence patterns categorized into High, Moderate, and Low categories.

112

Table 5. 3. Non-linear regression model used to model fish recovery as a function of downstream distance from barriers. Total and endemic fish recovery was characterized as rising immediately after a minimum distance threshold, as predicted consistently by the asymptotic exponential models. The chosen models are highlighted in bold.

Response Models Parameter estimates Residual Standard AIC variable compared Error, df Total Asymptotic a 0.57 (SE 0.06); p < 0.0001* 0.16, df = 9 -4.81 species exponential b 2.63 (SE 0.64); p = 0.0025* recovery Logistic a 0.81 (SE 1.85); p = 0.67 0.18, df = 8 -2.51 b -3.03 (SE 27.15); p = 0.91 c 7.90 (SE 43.94); p = 0.86 Endemic Asymptotic a 0.59 (SE 0.08); p < 0.0001* 0.21, df = 9 1.24 species exponential b 2.33 (SE 0.74); p < 0.012* recovery Logistic a 0.65 (SE 0.29); p = 0.05 0.25, df = 8 4.90 b -1.95 (SE 5.56); p = 0.74 c 2.29 (SE 7.53); p = 0.77

Note: Total and endemic fish recovery was characterized as rising immediately after a minimum distance threshold, as predicted consistently by the asymptotic exponential models. The chosen models are highlighted in bold. AIC = Akaike Information Criteria; *Statistical significance level alpha = 0.05

Discussion The potential of undammed tributaries to mitigate river regulation impacts These results suggest that recovery of freshwater fish species, including endemics from the WG, is still promising under the current level of hydrological regulation in the upper catchment of the Malaprabha river of the WG. Clearly, this result owes to the fact that undammed tributaries with adequate discharge (stream order 3) are still present in the basin, and appear to be mitigating the cumulative impacts of large and small barriers in the basin to some extent. Nearly all endemic species reappeared within a 5 km threshold distance downstream of barriers suggesting promising potential for recovery. Undammed tributaries might have high diversity of riverine habitats and serve as potential breeding grounds for fishes or as alternative migratory routes for potamodromous species (Sedell et al., 1990; King et al., 2009). It is likely that smaller barriers still offer partial or near-complete passage to fishes, especially during the monsoonal flooding pulse, as compared to the largest barrier in the study area. Our study highlights the contribution of undammed tributaries in replenishing catchment-scale functional connectivity across regulated stream networks (Johnson,

113

2002). This result is important for freshwater conservation planning to mitigate cumulative biodiversity impacts of river regulation in human-modified basins (Nunes et al., 2015).

Influences of local stream characteristics on species recovery Previous studies from regulated temperate and tropical river systems have reported partial recovery in terms of fish community composition (Niemi et al., 1990; Storey et al., 1991; Detenbeck et al., 1992). Proximity to reference sites and distance from barriers (Davey and Kelly, 2007; Kubach et al., 2011) were identified as key determinants of species recovery, as in our study. Other than distance, the nature of dam operations, age of barriers, local habitat conditions and anthropogenic impacts could aid or limit recovery (Ellis and Jones, 2013; Stoll et al., 2014; Piller and Gehber, 2015). For instance, the recovery of some species might have been influenced because of sensitivity to local pollution sources, sand-gravel mining or fishing impacts (e.g. rheophilic carps such as Tor khudree and Hypselobarbus dobsoni). The loach Paracanthocobitis mooreh, which requires sand-gravel substrates for spawning showed low recovery. Specialized habitat preferences of torrent-dwelling fishes such as Bhavania, Nemacheilus (loaches) and Glyptothorax (catfish) can lower the chances for their recovery. These species might have been locally extirpated or declined in abundance in the basin following regulation (authors; field observations), owing to their limited swimming and dispersal ability to escape habitat alterations caused by flow regulation (Daniels, 2002; Raghavan et al., 2008). Importantly, these species are either highly localized endemics or are enlisted as threatened (Molur et al., 2011) and hence the critical threshold of regulation for their recovery might not be identified without analyzing relative abundance patterns. For instance, the endemic carp Hypselobarbus dobsoni, which was feared to have disappeared from its native range in the WG (M. Arunachalam, pers. comm.) was detected, but only 5 km below barriers. At the same time, the presence of localized deep pools can provide refuge for species during the dry-season and drought periods and allow recovery in spite of barriers (Sedell et al., 1990) In turn, regulation may have benefitted limnophilic and omnivorous species such as Mystus catfishes, as seen also in the Bhadra river (V. Atkore, unpublished data) Fish communities that showed high recovery were dominated by rheophilic, heterotrophic, and periodic-breeding species, with whom

114 recovery levels showed no clear association. Interestingly, some benthic gravel- spawning catfishes, which we expected to be unaffected by regulation, showed low recovery.

For a more robust assessment of recovery for rare species, abundance data collected using methods to assess detectability and movement patterns, need to be incorporated in future assessments (Albanese et al., 2009). In addition, spatial variation in environmental factors (e.g. water quality) could have influenced recovery processes (Storey et al., 1991; Detenbeck et al., 1992). As noted in our results, dissolved oxygen increased and total alkalinity decreased with distance from barriers. Immediately downstream of barriers, groundwater contributions to stream flow would be higher, contributing to higher alkalinity, and at greater distances surface flows from undammed tributaries might restore surface- to groundwater ratios, lowering alkalinity. Unlike DO and alkalinity, the effects of water temperature and rocky habitat were independent of distance from barriers. Local human disturbances (fishing pressure, diversions for irrigation, water pollution, and substrate extraction) could have affected these variables and therefore are likely to affect species recovery irrespective of the influence of upstream barriers. Future detailed studies must attempt to understand how local anthropogenic disturbances and basin-scale flow alterations might affect recovery.

Simple metrics for rapid assessment of fish recovery in regulated river basins I believe that defining a reference species composition representative of unregulated conditions helped us perform neat comparisons of upstream unaffected sites with downstream sites under the impact of regulation (Voelz and Ward, 1989; Kubach et al., 2011). Detecting species recovery is challenging, given that pre-dam information on fish community composition is limited, or lacking entirely. As a result, space-for- time assessments of river regulation impacts are the norm in most regions, despite their limitations in detecting temporal changes in compositional patterns in freshwater taxa (Detenbeck et al., 1992). Ideally, long-term studies would be robust in detecting impacts of hydrological barriers through nested and intensive sampling schemes, to understand processes of species recovery (Ryon, 2011; Ellis and Jones, 2013), subject to broader influences of climate variability and land-use changes (Dee Boersma et al.,

115

2001; Kibler and Tullos, 2013). Future studies must also assess fish movements both upstream and downstream of barriers using mark-recapture framework (Albanese et al., 2003). Though this study is limited in scope to capture these dynamics, it provides simple and effective metrics of fish community recovery that may find application in rapid ecological assessments (Galatowitsch et al., 1998).

Implications for river restoration and conservation in tropical rivers To maintain high recovery in dammed basins, my observations suggest that 5-10 undammed and undisturbed 2nd and 3rd order streams each need to be protected in the catchment of approximately 500 sq.km area. The observed recovery might indicate a tipping point beyond which irreversible losses (local extinctions) are likely if remaining tributaries are also slated for future regulation. Flows of these remnant streams are already being abstracted for intensive agriculture. These results indicate that strictly limiting any future regulation of these streams will be critical to sustain fish species recovery. Smaller barriers might have low impact on species recovery and their impacts in future water development plans need to be assessed case-by-case through detailed field-based ecological flow modelling exercises to identify thresholds needed for effective recovery. The recovery thresholds (2-5 km) that I report can be useful indicators of cumulative impacts on recovery perhaps applicable to other regulated river basins in the WG. Similar levels of regulation and protection were thought to enhance population persistence of endemic and threatened fish species in the regulated Muse and Dutch rivers (Leeuw et al., 2005).

A recent inter-basin transfer from the headwaters of the Mhadei to the east-flowing Malaprabha basin could potentially affect river habitats essential for specialized endemic fishes (Atkore et al., 2012). The canal plans to divert 7.5 thousand cubic meter (tmc) of water annually to two cities located downstream in Karnataka state for drinking water purpose. Recently, the Mhadei Water Dispute Tribunal in its interim order on 27th July 2016 rejected Karnataka’s demand for water which led to escalated protests in many cities of Karnataka. These protests make it clear that human appropriation of river water is only going to increase in this basin, and adverse impacts on river fish diversity may be imminent with continued flow regulation. In this context, our results can inform policy-makers on how conservation of stream

116 fishes and other riverine biodiversity could be achieved, despite intensifying demands, in future negotiations on water sharing and management in this basin.

References

Abraham, Robin Kurian, and Kelkar, Nachiket. 2012. Do terrestrial protected areas conserve freshwater fish diversity? Results from the Western Ghats of India. Oryx 46(04), 544-553. doi: doi:10.1017/S0030605311000937

Albanese, B., Angermeier, P. L., and Gowan, C. 2003. Designing Mark–Recapture Studies to Reduce Effects of Distance Weighting on Movement Distance Distributions of Stream Fishes. Transactions of the American Fisheries Society, 132(5), 925-939. doi: 10.1577/T03-019

Albanese, Brett, Angermeier, Paul L., and Peterson, James T. 2009. Does mobility explain variation in colonisation and population recovery among stream fishes? Freshwater Biology 54(7), 1444-1460. doi: 10.1111/j.1365- 2427.2009.02194.x

Alexandre, C. M., Ferreira, T. F., and Almeida, P. R. 2013. Fish assemblages in non- regulated and regulated rivers from permanent and temporary Iberian systems. River Research and Applications 29, 1042-1058.

Anderson, Elizabeth P., Freeman, Mary C., and Pringle, Catherine M. 2006. Ecological consequences of hydropower development in Central America: impacts of small dams and water diversion on neotropical stream fish assemblages. River Research and Applications 22(4), 397-411. doi: 10.1002/rra.899

APHA. 2005. Standard Methods for Examination of Water and. Wastewater, American Public Health Association WWA,. Washington, D.C. Pp. 1496.

Atkore, V., J. Krishnaswamy., S. Badiger., and Shanker, K. 2012. Preliminary assessment of fish fauna in the upper Malaprabha river basin of the Western Ghats, India. Paper presented at the 4th Bio-Asia Regional Seminar, Bio-Asia Program, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China.

Bhat, A., and Magurran, A. E. 2007. Does disturbance affect the structure of tropical

117

fish assemblages? A test using null models. Journal of Fish Biology 70(2), 623-629. doi: 10.1111/j.1095-8649.2006.01252.x

Brown, J. Jed, Limburg, Karin E., Waldman, John R., Stephenson, Kurt, Glenn, Edward P., Juanes, Francis, and Jordaan, Adrian. 2013. Fish and hydropower on the U.S. Atlantic coast: failed fisheries policies from half-way technologies. Conservation Letters 6(4), 280-286. doi: 10.1111/conl.12000

Brown, Larry R., and Ford, Tim. 2002. Effects of flow on the fish communities of a regulated California river: implications for managing native fishes. River Research and Applications 18(4), 331-342. doi: 10.1002/rra.673

Connell, Joseph H., and Sousa, Wayne P. 1983. On the Evidence Needed to Judge Ecological Stability or Persistence. The American Naturalist 121(6), 789-824.

Crawley, M. J. 2007. The R Book. John & Wiley Sons, Ltd. The Atrium Southern Gate, Chichester, West Sussex PO 19 8SQ, England. Pp. 877.

Dahanukar, N, and Raghavan, R. 2013. Freshwater fishes of the Western Ghats: Checklist v1.0 August 2013. Min. 01, Pp. 6-16.

Daniels, R.J.R. 2002. Freshwater fishes of Peninsular India. Hyderabad: University Press India (Pvt) Ltd. Pp.321.

Davey, Andrew J. H., and Kelly, David J. 2007. Fish community responses to drying disturbances in an intermittent stream: a landscape perspective. Freshwater Biology 52(9), 1719-1733. doi: 10.1111/j.1365-2427.2007.01800.x

Dee Boersma, P., Kareiva, Peter, Fagan, William F., Alan Clark, J., and Hoekstra, Jonathan M. 2001. How Good Are Endangered Species Recovery Plans? BioScience 51(8), 643-649. doi: 10.1641/0006- 3568(2001)051[0643:HGAESR]2.0.CO;2

Detenbeck, Naomi E, DeVore, PhilipW, Niemi, Gerald J, and Lima, Ann. 1992. Recovery of temperate-stream fish communities from disturbance: A review of case studies and synthesis of theory. Environmental Management 16(1), 33-53. doi: 10.1007/BF02393907

Dudgeon, D. 2011. Asian river fishes in the Anthropocene: threats and conservation challenges in an era of rapid environmental change. Journal of Fish Biology 79(6), 1487-1524. doi: 10.1111/j.1095-8649.2011.03086.x

Dudgeon, David. 2010. Prospects for sustaining freshwater biodiversity in the 21st century: linking ecosystem structure and function. Current Opinion in Environmental Sustainability. 2(5–6), 422-430. doi: http://dx.doi.org/10.1016/j.cosust.2010.09.001

118

Ellis, Lucy Eunsun, and Jones, Nicholas Edward. 2013. Longitudinal trends in regulated rivers: a review and synthesis within the context of the serial discontinuity concept. Environmental Reviews. 21(3), Pp. 136-148. doi: 10.1139/er-2012-0064

Galatowitsch, Susan M., Whited, Diane C., and Tester, John R. 1998. Development of community metrics to evaluate recovery of Minnesota wetlands. Journal of Aquatic Ecosystem Stress and Recovery 6(3), 217-234. doi: 10.1023/a:1009935402572

Hoeinghaus, David J., Agostinho, Angelo A., Gomes, Luiz C., Pelicice, Fernando M., Okada, Edson K., Latini, JoÃO D., and Winemiller, Kirk O. 2009. Effects of River Impoundment on Ecosystem Services of Large Tropical Rivers: Embodied Energy and Market Value of Artisanal Fisheries. Conservation Biology 23(5), 1222-1231. doi: 10.1111/j.1523-1739.2009.01248.x

Ibañez, Carla, Belliard, Jérôme, Hughes, Robert M., Irz, Pascal, Kamdem-Toham, André, Lamouroux, Nicolas, Oberdorff, Thierry. 2009. Convergence of temperate and tropical stream fish assemblages. Ecography 32(4), 658-670. doi: 10.1111/j.1600-0587.2008.05591.x

Jackson, Donald A., Peres-Neto, Pedro R., and Olden, Julian D. 2001. What controls who is where in freshwater fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries and Aquatic Sciences 58(1), 157- 170. doi: 10.1139/f00-239

Jha, R., and Singh, Vijay P. 2008. Evaluation of riverwater quality by entropy. Water Engineering 12(1), 61-69.

Johnson, C. W. 2002. Riparian vegetation diversity along regulated rivers: contribution of novel and relict habitats. Freshwater Biology 47: 749-759. DOI: 10.1046/j.1365-2427.2002.00910.x.

Kibler, M., Kelly, and Tullos, Desiree D. 2013. Cumulative biophysical impact of small and large hydropower development in Nu River, China. Water Resources Research 49(6), 3104-3118. doi: 10.1002/wrcr.20243

King, A. J., Tonkin, Z., and Mahoney, J. 2009. Environmental flow enhances native fish spawning and recruitment in the Murray River, Australia. River Research and Applications 25(10), 1205-1218. doi: 10.1002/rra.1209

King, Alison J., Gwinn, Daniel C., Tonkin, Zeb, Mahoney, John, Raymond, Scott, and Beesley, Leah. 2015. Using abiotic drivers of fish spawning to inform environmental flow management. Journal of Applied Ecology DOI: 10.1111/1365-2664.12542.

Kubach, Kevin M., Scott, Mark C., and Bulak, James S. 2011. Recovery of a temperate riverine fish assemblage from a major diesel oil spill. Freshwater Biology 56(3), 503-518. doi: 10.1111/j.1365-2427.2010.02517.x

119

Leeuw, J. J. de, Buijse, A. D., Grift, R. E., and Winter, H. V. 2005. Management and monitoring of the return of riverine fish species following rehabilitation of Dutch rivers. Fundamental and Applied Limnology / Archiv für Hydrobiologie, 155(Suppl. 1-4), 391-411.

Macnaughton, Camille J., McLaughlin, Fraser, Bourque, Guillaume, Senay, Caroline, Lanthier, Gabriel, Harvey-Lavoie, Simonne, and Boisclair, Daniel. 2015. The Effects of Regional Hydrologic Alteration on Fish Community Structure in Regulated Rivers. River Research and Applications DOI: 10.1002/rra.2991. doi: 10.1002/rra.2991

Malmqvist, Bjorn , and Rundle, Simon. 2002. Threats to the running water ecosystems of the world. Environmental Conservation 29 (02), 134-153. doi: doi:10.1017/S0376892902000097

Molur, S., K. G. Smith, Daniel, B. A., and Darwall, W. R. T. 2011. The Status and distribution of freshwater biodiversity in the Western Ghats, India. Cambridge, UK and Gland, Switzerland: IUCN, and Coimbatore, India: Zoo Outreach Organisation. (pp. 128): IUCN.

Niemi, G. J., DeVore, P., Detenbeck, N., Taylor, D., Lima, A., Pastor, J., Yount, J. David, Naiman, R.J. 1990. Overview of case studies on recovery of aquatic systems from disturbance. Environmental Management 14(5), 571-587. doi: 10.1007/bf02394710

Nilsson, Christer, Reidy, Catherine A., Dynesius, Mats, and Revenga, Carmen. 2005. Fragmentation and Flow Regulation of the World's Large River Systems. Science, 308(5720), 405-408. doi: 10.1126/science.1107887

Nunes D. M. F., Magalhães A. L. B., Weber A. A., Gomes R, Normando F. T., Santiago K.B., Rizzo E, and Bazzoli N. 2015. Influence of a large dam and importance of an undammed tributary on the reproductive ecology of the threatened fish matrinxã Brycon orthotaenia Günther, 1864 (Characiformes: Bryconidae) in southeastern Brazil. Neotropical Ichthyology 13:317-324. DOI: 10.1590/1982-0224-20140084.

Phomikong, P., Fukushima, M., Sricharoendham, B., Nohara, S., and Jutagate, T. 2015. Diversity and Community Structure of ivFishes in the Regulated Versus Unregulated Tributaries of the Mekong River. Rer Research and Applications 31(10), 1262-1275. doi: 10.1002/rra.2816

Piller, Kyle R., and Geheber, Aaron D. 2015. Black liquor and the hangover effect: fish assemblage recovery dynamics following a pulse disturbance. Ecology and Evolution 5(12), 2433-2444. doi: 10.1002/ece3.1530

120

Pringle, C. M. 2001. Hydrologic connectivity and the management of biological reserves: a global perspective. Ecological Application 11(4), 981-998.

R core Team 2013. R: A language and environment for statitical computing. R Foundation for statistical computing, Vienna, Austria. URL:http://www.R- project.org/

Rema Devi, K., Indra, T. J., Yadav, B.E., Raghunathan, M. B., Krishnan, S. and Jadhav, S. S. 2013. Freshwater fishes In 'State Fauna Series. Fauna of Karnataka' In Zool. Surv. India Editor-Director (Ed.), State Fauna Series (Vol. 21, pp. 1-595). Kolkata: Zoological Survey of India.

Rolls, R. J., and Sternberg, D. 2015. Can Species Traits Predict the Susceptibility of Riverine Fish to Water Resource Development? An Australian Case Study. Environmental Management, 55(6), 1315-1326. doi: 10.1007/s00267-015- 0462-8

Ryon, Michael G. 2011. Recovery of Fish Communities in a Warm Water Stream Following Pollution Abatement. Environmental Management 47(6), 1096- 1111. doi: 10.1007/s00267-010-9596-x

Sakaris, P. C. 2013. A review of the effects of hydrologic alteration on fisheries and biodiversity and the management and conservation of natural resources in regulated river systems.

Scheidegger, Karl J., and Bain, Mark B. 1995. Larval Fish Distribution and Microhabitat Use in Free-Flowing and Regulated Rivers. Copeia 1995(1), 125-135. doi: 10.2307/1446807

Schlosser, Isaac J. 1991. Stream Fish Ecology: A Landscape Perspective. BioScience 41(10), 704-712. doi: 10.2307/1311765

Sedell, James R., Reeves, Gordon H., Hauer, F. Richard, Stanford, Jack A., and Hawkins, Charles P. 1990. Role of refugia in recovery from disturbances: Modern fragmented and disconnected river systems. Environmental Management 14 (5), 711-724. doi: 10.1007/bf02394720

Stanford, J. A., and Ward, J. V. 2001. Revisiting the serial discontinuity concept. Regulated Rivers: Research & Management 17(4-5), 303-310. doi: 10.1002/rrr.659

Stoll, Stefan, Kail, Jochem, Lorenz, Armin W., Sundermann, Andrea, and Haase, Peter. 2014. The Importance of the Regional Species Pool, Ecological Species Traits and Local Habitat Conditions for the Colonization of Restored River Reaches by Fish. PLoS ONE 9(1), e84741. doi: 10.1371/journal.pone.0084741

121

Storey, A. W., Edward, D. H., and Gazey, P. 1991. Recovery of aquatic macroinvertebrate assemblages downstream of the Canning Dam, Western Australia. Regulated Rivers: Research and Management 6(3), 213-224. doi: 10.1002/rrr.3450060306

Taylor, Jason M., King, Ryan S., Pease, Allison A., and Winemiller, Kirk O. 2014. Nonlinear response of stream ecosystem structure to low-level phosphorus enrichment. Freshwater Biology 59(5), 969-984. doi: 10.1111/fwb.12320

Theophilus, Emmanuel. 2014. Headwater Extinctions- Hydropower projects in the Himalayan reaches of the Ganga and the Beas: A closer look at impacts on fish and river ecosystems (pp. 80). New Delhi: SANDRP.

Thorp, J. H., Martin C. Thoms, and Michael, D. D. 2008. The riverine ecosystem synthesis: Towards Conceptual Cohesiveness in River Science. Academic Press. Pp. 223

Travnichek, Vincent H., Bain, Mark B., and Maceina, Michael J. 1995. Recovery of a Warmwater Fish Assemblage after the Initiation of a Minimum-Flow Release Downstream from a Hydroelectric Dam. Transactions of the American Fisheries Society 124(6), 836-844. doi: 10.1577/1548- 8659(1995)124<0836:ROAWFA>2.3.CO;2

Vannote, R. L., Minshall, G. W., Cummins, K. W., Sedell, J. R., and Cushing, C. E. 1980. The River Continuum Concept. Canadian Journal of Fisheries and Aquatic Sciences 37(1), 130-137. doi: 10.1139/f80-017

Voelz, Neal J., and Ward, J. V. 1989. Biotic and abiotic gradients in a regulated high elevation Rocky mountain river. Regulated Rivers: Research and Management 3(1), 143-152. doi: 10.1002/rrr.3450030114

Vorosmarty, C. J., McIntyre, P. B., Gessner, M. O., Dudgeon, D., Prusevich, A., Green, P., . . . Davies, P. M. 2010. Global threats to human water security and river biodiversity. Nature, 467(7315), 555-561. doi: http://www.nature.com/nature/journal/v467/n7315/abs/nature09440.html#supp lementary-information

Ward, J. V., and Stanford, J. A. 1995. The serial discontinuity concept: Extending the model to floodplain rivers. Regulated Rivers: Research and Management 10(2- 4), 159-168. doi: 10.1002/rrr.3450100211

Welcomme, R. L., Winemiller, K. O., and Cowx, I. G. 2006. Fish environmental guilds as a tool for assessment of ecological condition of rivers. River Research and Applications 22(3), 377-396. doi: 10.1002/rra.914

122

Chapter 6

Synthesis and Conclusion

Globally, freshwater ecosystems are severely threatened due to over-exploitation both for water and energy demands of society. As a result, freshwater fish – one of the important components of these ecosystems - have become highly vulnerable. Stream fish community ecology has been well studied both in the temperate and tropical regions. Typically, most studies conducted have evaluated the patterns of diversity, abundance and distribution which often interest stream ecologists. However, due to increasing threats to freshwater systems, extinction risk for riverine fishes is exceeding that of the natural rate of extinction (Dias et al., 2017). Studies that quantify fish-environment relationship are on the rise (Latin America, Mediterranean, Australia, and New Zealand) including Asian rivers (South and South East Asia).

Tropical Asian rivers systems harbor a wide array of habitats that support extraordinarily rich aquatic diversity (Dudgeon, 2000). India hosts more than 950 freshwater fish species in its diverse freshwater habitats (lakes, streams and rivers) (Jayaram, 2010). However, these freshwater habitats are being over exploited for various developmental projects and water diversion schemes.

The WG region, part of the Western Ghats–Sri Lanka Biodiversity hotspot - currently holds particularly high freshwater fish endemic diversity (66 %) compared to other freshwater taxa (Dahanukar & Raghavan, 2013; Raghavan et al., 2016). IUCN reports suggest that our knowledge of patterns of diversity and distribution of fish communities are still incomplete in this region (Molur et al., 2011). However, basin- wide approaches within a larger ecoregion which enable an understanding of the complete structure and organization of fish communities at multiple spatial scales ranging from local (segment) to regional (stream order) or basin level (catchment level), are lacking (Hauer and Lamberti, 2007). This can also help disentangle anthropogenic and environmental factors that drive species composition across scales. Thus data collected in a nested design within and across basins offer a consistent surveying methodology enabling both inter and intra river comparison in fish

123 communities with regard to its stream environment (Macnaughton, 2016). Previous studies conducted in this region on fish communities have contributed to knowledge on fish-stream characteristics (stream order, habitat) and few fish guilds (Arunachalam, 2000; Bhat, 2004; Raghavan et al., 2008; Johnson and Arunachalam, 2010; Abraham and Kelkar, 2012; Chakrabarty and Homechaudhuri, 2013; Kundu et al., 2014) but did not include water chemistry, comprehensive fish guilds or assess the effect of anthropogenic or environmental factors on fish guilds. None of these studies have attempted to understand the possibility of species recovery especially below the dam in changing riverscapes.

The primary objective of my thesis was to understand the different drivers of fish diversity (richness, abundance) and composition within as well as across four river sub-basins and evaluate the possibility of species recovery below hydrological barriers. Overall, I recorded a total of 93 species belonging to nine orders and 18 families from 18,322 individuals. About 99% of the diversity was comprised of three orders – Cypriniformes, Siluriformes and Perciformes. Of the 93 species sampled, I found about 50 fish species were endemic to the WG region (Molur et al., 2011; Dahanukar and Raghavan, 2013). Some genera such as Barilius and Barbodes were restricted only to the relatively unregulated sub-basins (Bhadra and Tunga) while other genera such as Sicyopterus and Mugil were found only in regulated west- flowing Mhadei basin. Malaprabha had three genera exclusively to its basin. They include Chela, Systomus and Schimatorhynchus. The distribution of these genera in each of the basins suggest a biogeographical effect.

I attempted to understand factors (environmental and anthropological) that drive the native fish richness and abundance in 152 stream segments spread across four sub- basins. Several important insights emerge from this study that would potentially aid river conservation efforts both for the rivers in the WG’s, and elsewhere in India: (1). Species diversity (richness and abundance) was higher in large rivers (Mhadei and Malaprabha) than in smaller rivers (Tunga and Bhadra); (2) River regulation impacted surface dwelling guilds while benefiting mid and bottom dwelling fish guilds; (3) Water chemistry variables such as dissolved oxygen, calcium hardness and alkalinity also influence diverse fish guilds indicating importance of water monitoring; (4) Non- regulated rivers harbor diverse stream substratum that in turn shape fish community

124 composition; (5) fish species recovery can be achieved even in the regulated basins provided undammed tributaries are maintained below existing barriers. In second chapter, I examined the drivers of species richness and abundance in four river sub-basins at multiple spatial scales. Information on site specific richness, abundance and species composition and habitat use of stream fish communities is vital for river conservation. My results based on data collected across 152 stream segments on environmental and fish diversity suggest that fish diversity (richness and abundance) was influenced by diverse set of stream characteristics as well as water chemistry. My results suggest that diverse fish species have a complex association with environment and water chemistry. Streams with high disturbance levels were associated with higher species richness while fish abundance was higher in less disturbed segments, indicating that generalist and disturbance tolerant species contributed to greater richness in the WG. However, WG endemic and generalist species also utilized less disturbed environments in the spawning period.

In terms of river basins, Malaprabha was the most speciose with 53 species, of the three sub-basins (Mhadei, Tunga and Bhadra). The presence of a reservoir downstream of the Malaprabha river facilitated fish movement upstream resulting in a higher observed species richness. Additionally, pool habitat modified by various impoundment structures (check dams, barrages) have created a niche space allowing many species to coexist. Mhadei was the second richest with 47 species and this could be due to the presence of a good network of perennial streams of 1st and 2nd orders apart from the habitat modification due to many river barriers (check dams), that again might lead to increased niches. Tunga sub-basin was the third sub-basin in species richness with 45 species. Species richness in Tunga was possibly due to the presence of perennial streams of 1st, 2nd and 3rd stream orders that provide a gentle gradient in the sub-basin, which in turn yields a variety of river habitats (run, riffle and pool). Bhadra river was very similar to Tunga in terms of its habitat diversity and perennial stream network apart from also being connected to a reservoir downstream. However, it was relatively species poor with only 24 species compared to other sub- basins. A steep gradient within the basin due to natural waterfalls situated in the headwater region could have restricted fish movement. The habitat heterogeneity hypothesis suggests that more diverse habitats support higher species richness (Ricklefs and Dolph, 1993). However, this study suggests that the homogeneous

125 habitats (pool) formed due to temporary impoundments as well as small check dams supported a higher species richness.

I also evaluated the conservation status of all the species sampled in the study area according to the current IUCN Red list (Molur et al., 2011). I found that 16 species identified in the study were not evaluated previously by IUCN. Additionally I described a new fish species called Kudremukh barb (Pethia striata) from Tunga sub- basin. These findings indicate that more systematic survey efforts are required to discover new fish species from the headwater regions (Atkore et al., 2015). I also report a healthy population of critically endangered species Barbodes wynaadensis within the Kudremukh National Park and a record of Hypselobarbus thomassi in the Malaprabha sub-basin. A new extension record of Labeo dussumieri from Bhadra river is reported from this study. Moreover, there is an urgent need to focus scientific and conservation attention on not just species but also their potential breeding habitats.

Species diversity (richness) measures alone do not reveal true community composition which is often important for conserving complex river systems in human dominated landscapes (McKnight, 2007). Human impacts in stream environment resulted either species gain (via introduction) or species loss (extinction) (Xu et al., 2015), and therefore, assessing the species composition or species turnover in relatively less disturbed river systems offer vital ecological insights as it further guide us prioritize crucial river segments for their protection. The Bhadra and Tunga sub- basins (relatively unregulated basins) were studied to determine species turnover (changes in species composition along a river gradient). Both of these rivers originate from the same geographical area (Gangamoola at 1160 m elevation in Kudremukh National Park) and share similar environmental features (rainfall, temperature, habitat and land use types). The species turnover or spatial structure of fish communities was studied at two scales, at local (segment) and at a sub-basin scales. Mantel’s correlogram indicated a sine-hole pattern suggesting a high turnover at smaller spatial scale in both the sub-basin which also indicates dispersal limitation (Astorga et al., 2011; 2014). Mantel’s r was also positively correlated with stream order, canopy cover, and negatively correlated with width to depth ratio and water temperature in Bhadra basin, while in Tunga, Mantel’s r was positively correlated with substrate and

126 negatively correlated with electrical conductivity. This strong influence of environmental variables on species turnover suggests that niche based processes may play an important role in structuring of these communities (Astorga et al., 2011; 2014).

These results clearly suggest that unregulated river segments especially situated in the headstream regions (within the national park boundary) offer habitat for species composition. A gentle elevation gradient maintain diverse habitat structure, facilitates fish mobility upstream-downstream environment, good water quality and species composition. Certain habitat specialist fish species only prefer relatively undisturbed environment to complete their life cycle (e.g. Tor khudree, Barbodes wynaadensis, Balitora mysorensis etc found in these basins) and any habitat alteration might severely affect their survival in the future.

Species richness and composition yield rich ecological knowledge on aquatic communities but, information on species guilds/ traits reveal functional aspect of community is often neglected from the Asian tropical river systems. To understand how hydrological barriers influence various fish guilds in two regulated and two non- regulated river sub-basin, I classified all sampled fish species that share similar geography (stream order, elevation etc) into different fish guilds based on primary data (intensive field data collection) as well as secondary data gathered from grey literature including theses, reports, books and published work (Lowe-McConnell, 1975; 1987; De Silva et al., 1979; Welcomme, 1985; Bhat, 2002; Johnson and Arunachalam, 2012; Chakrabarty and Homechaudhuri, 2013). Guild species composition varied between regulated and non-regulated sub-basin with more than half of the endemic species (including headwater habitat specialist) confined to non- regulated sub-basin. To determine, whether hydrological degradation was also due to excessive water abstraction, I analyzed data on disturbance regime as well as rainfall and river discharge data collected for the period 1979 to 2013 from the state irrigation departments of Goa and Karnataka. I performed time series analysis to understand the trend in rainfall and discharge over these years. The rainfall showed insignificant decline but discharge declined significantly.

Higher levels of calcium hardness and total dissolved solids indicated poor water

127 quality. Responses to environmental variables differed across three guilds. Overall, calcium hardness was negatively and total dissolved solids was positively correlated with guild richness. Additionally, guilds richness was higher for two of three guilds in regulated sub-basin than in the non-regulated sub-basin. Of the three guilds, the surface dwelling guild was negatively affected by river barriers (blocking their movement) while mid and bottom dwelling guilds benefited, likely due to an impoundment effect (Aarts and Nienhuis, 2003; Kanno and Vokoun, 2010). Changing water chemistry and river barriers together affect water column based fish guilds in the WG region. Future studies need to incorporate data on guild abundance, and water abstraction across many comparable river sub-basins. The insights from these results suggest that often neglected data on functional guilds might be crucial in forming policy in re-designing hydrological barriers in tropical region or designing mitigation measures in regulated rivers.

So far, studies on fish communities have demonstrated the factors causing fish diversity at multiple spatial scales in different ecoregions (Matthews, 1998; Pusey et al., 1998; Bhat, 2003; Gido and Jackson, 2010; Arthington et al., 2014). Given that rivers are heavily appropriated worldwide for water and energy demands, very few studies have actually attempted to understand species recovery downstream of a dam. In a hydrologically modified river basin (such as Malaprabha), I evaluated the role of species recovery with distance downstream below the dam. I hypothesized that, species recovery will decrease immediately below the dam and increase with increasing distance from the dam due to the contribution of the undammed tributaries. I tested and validated this hypothesis, and my results show that total species similarity (proportion of species encountered in dammed river segments to undammed river segments) and endemic species similarity (recovery) was low immediately downstream of a dam but increased with increasing distance from the dam. Species recovery was higher within a distance of 2 km and it reduced beyond 5 km. The pattern was also similar in the case of endemic species recovery. This recovery below the dam was mainly attributed to the joining of undammed tributaries to the main river channel which not only provides a refuge, feeding and breeding ground for many fish species but also ameliorates river water quality due to instream flow (Ellis and Jones, 2013; Penczak et al., 2014; Espirito-Santo and Zuanon, 2016). Future studies must assess the fish movement from upstream to downstream of barriers using mark

128 recapture frameworks (Albanese et al., 2003) and collect more data on fish abundance and associated life-history characteristics to understand the role of species recovery. Although, I found fairly good evidence of species recovery in a regulated sub-basins of the WG’s, this recovery could be temporary, given that most of the undammed and undisturbed tributaries in the river basin are being exploited for agriculture as well as hydropower regulation in near future. If such interventions are continued, fish decline may continue in the downstream sections and leading to low species recovery.

Conservation implications This study is the first of its kind focused on freshwater fish which involve rigorous field sampling across four sub-basins of the WG rivers. The study compares the combined influence of hydrological barriers, water chemistry, environmental factors on diverse fish guilds, and assess the species recovery below the dam. Given the dwindling nature of freshwater habitat in the country, results emerging from this study have a huge conservation science implications. Similar studies can be replicated in other parts of the country to answer key ecological questions that may highlight the ecological as well as conservation significance of last free-flowing rivers in India. It is necessary to monitor water quality variables along with fish diversity in both disturbed and non-disturbed rivers. Composition of fish guilds may serve as important ecological indicators to understand the health of river systems. Conservation prioritization should therefore identify continuous stream segments that harbors diverse habitats with good water quality. It is also important to maintain few free flowing river sections that allow free movement for river fishes to perform feeding and breeding movements.

Overall, we need to prioritize undammed tributaries for protection from local pollution and monitor water chemistry in local hotspots of endemic fish diversity. Data generated from this study has filled a current knowledge gap on distribution as well as on the population status of many threatened species.

River segments which are less modified and perennial in nature offer diverse habitat types and substrate composition which are essential for maintaining headwater specialists and overall species composition. Maintaining a good network of connected

129 stream orders in a basin provides a healthy fish population as well as an opportunity to test the structure of stream fish communities. Species turnover seems to be driven mainly by stream substrate and elevation (600-800 m) at short spatial scales. It is important to prioritize this elevation range for maintaining species composition within protected areas.

Diversity of fish in terms of their morphology, behavior and reproductive strategies offer a unique opportunity to understand how functional guilds provide an intuitive understanding of fish communities (Mouillot et al., 2013) and therefore, studying functional traits/guilds is an exciting field of research in stream fish community ecology (Winemiller et al., 2010). Among the various guild types, guilds such as water-column based guilds seem to provide interesting insights in relation to environmental variables as well as hydrological barriers. Future long-term monitoring programs should incorporate guild level information. Although crucial information on life-history strategies and breeding biology of many Indian fishes are grossly lacking, the guild-level information that has been generated as part of this work will aid in ascertaining future conservation strategies for rivers in the WG’s as well as other Indian rivers.

It is important to separate effects of hydrological barriers, water chemistry and environmental variables and assess how fish composition is influenced by each factor and in combination. It is also important to quantify trade-off between hydrological regulation and environmental variables in shaping fish community composition. This study provides useful insights on some of these aspects. Having long-term data on stream discharge is vital in order to draw meaningful insights into how fish communities are influenced in regulated and non-regulated river sections. Future studies are required to take into consideration a basin with a good stream network that provides both control as well as disturbed segments (dams and other anthropogenic disturbances such as water uptake, substrate mining, fishing etc). Based on precautionary principle invoked from the study, a basin with 500 sq. km area (based on the sub-basin area of Malaprabha river) should maintain at least five to ten undammed streams to test the idea of species recovery. Studies should also focus on other factors that are likely to influence species recovery i.e. longitudinal connectivity, diversity in stream habitats, good water quality, biological traits of fish species and

130 distance from the barrier etc.

In summary, I find that at smaller scales (segment level), stream fish communities (richness and abundance) are structured primarily by water chemistry and at larger spatial scales by stream size. At small spatial scales, fish species composition is influenced by substrate heterogeneity and at sub-basin scale, by stream order (surrogate for elevation). Fish guild richness was influenced primarily by water chemistry, environmental and hydrological barriers. However, this study did not assess the relationship between other climatic factors on fish communities at the basin or WG scale. It would be useful to incorporate such knowledge on native fish communities of the WG in India. Results from my thesis may provide useful insights for river management as the fish metric developed in the study (fish-environment relationship and fish recovery) might help in assessing the health of a river system. Systematic and comprehensive methodology developed in the study may aid in the river monitoring efforts. Identifying and maintaining reference sites in a modified river basin might also be useful metric to compare and assess the state of river degradation. This thesis answers some of the methodological and fundamental questions with regard to fish community structure in both modified and unmodified basins. And finally, results generated from my thesis may serve as a useful guide for future studies that aim to investigate the relationship between fish diversity indices and long term water chemistry and environment data.

References

Albanese, B., Angermeier, P. L., and Gowan, C. 2003. Designing mark‐recapture studies to reduce effects of distance weighting on movement distance distributions of stream fishes. Transactions of the American Fisheries Society, 132(5), 925–939. https://doi.org/10.1577/T03-019

Arthington, A. H., Rolls, R. J., Sternberg, D., Mackay, S. J., & James, C. S. 2014. Fish assemblages in subtropical rivers: low-flow hydrology dominates hydro- ecological relationships. Hydrological Sciences Journal, 59(3-4), 594-604. doi: 10.1080/02626667.2013.844345

131

Arunachalam, M. 2000. Assemblage structure of stream fishes in the Western Ghats (India). Hydrobiologia 430 (1-3), 1-31. doi: 10.1023/A:1004080829388

Astorga, Anna, Heino, Jani, Luoto, Miska, and Muotka, Timo. 2011. Freshwater biodiversity at regional extent: determinants of macroinvertebrate taxonomic richness in headwater streams. Ecography, 34 (5), 705-713. doi: 10.1111/j.1600-0587.2010.06427.x

Atkore, V. M., Knight, J. D. Marcus, Devi, K. Rema, and Krishnaswamy, J. 2015. A New Species of Pethia from the Western Ghats, India (Teleostei: Cyprinidae). Copeia 103 (2), 290-296. doi: 10.1643/OT-12-172

Bhat, A. 2002. A study of the diversity and ecology of freshwater ﬁshes of four river systems of the Uttara Kannada District, Karnataka, India. A PhD thesis submitted to the Centre for Ecological Scienes, Indian Institute of Science, Bangalore. Pp. 178.

Bhat, A. 2003. Diversity and Composition of Freshwater Fishes in River Systems of Central Western Ghats, India. Environmental Biology of Fishes 68 (1), 25-38. doi: 10.1023/a:1026017119070

Borcard, D., Legendre, P., & Drapeau, P. 1992. Partialling out the spatial component of ecological variation. Ecology 73. doi: 10.2307/1940179

Borcard, Daniel; , Gillet, Francois;, and Legendre, Pierre 2011. Numerical Ecology with R. Montréal, QC & Besançon, France: Springer.Pp. 319.

Dahanukar, N, and Raghavan, R. 2013. Freshwater fishes of the Western Ghats: Checklist v1.0 August 2013. Min 01, 6-16.

Dais, M. S., Pablo A. Tedesco, Bernard Hugueny, Celine Jezequel, Olivier Beauchard, Sebastien Brosse, and Thierry Oberdorff. 2017. Anthropogenic stressors and riverine extinctions. Ecological Indicators. doi:10.1016/j.ecoind.2017.03.053.

132

De Silva, S.S., Cumaranatunga, P.R.T., and De Silva, C.D. 1979. Food, Feeding Ecology and Morphological Features Associated With Feeding of Four Co- Occurring Cyprinids (Pisces: Cyprinidae). Netherlands Journal of Zoology 30 (1), 54-73. doi: doi:10.1163/002829680X00032

Dudgeon, David. 2000. The Ecology of Tropical Asian Rivers and Streams in Relation to Biodiversity Conservation. Annual Review of Ecology and Systematics 31, 239-263.

Espírito-Santo, Helder M. V., and Zuanon, Jansen. 2016. Temporary pools provide stability to fish assemblages in Amazon headwater streams. Ecology of Freshwater Fish n/a-n/a. doi: 10.1111/eff.12292

Gido, K., and Jackson, Donald A. 2010. Community Ecology of Stream Fishes: Concepts, Approaches, and Techniques. Bethseda, Maryland, USA: American Fisheries Society Sympposium 73. Pp.665.

Goslee, S and, and Urban, Dean. 2007. The ecodist package for dissimilarity-based analysis of ecological data. Journal of Statistical softwarre 22 (7), 1-19.

Hauer, F.R and, and Lamberti, G. 2007. Methods in stream ecology (Vol. 2nd): Academic Press.

Jayaram, K. C. 2010. The Frehswater Fishes of Indian Region. Delhi: Narendra Publishing House.

Johnson, J. A. and Arunachalam, Muthukumarasamy. 2010. Habitat use of fishes in streams of Kalakad Mundanthurai Tiger Reserve, India. International Journal of Ecology & Development, 17 (F 10), 34-47.

Johnson, J. A., and Arunachalam, Muthukumarasamy. 2012. Feeding habit and food partitioning in a stream fish community of Western Ghats, India. Environmental Biology of Fishes 93 (1), 51-60. doi: 10.1007/s10641-011- 9889-9.

Kanno, Y., and Vokoun, J. C. 2010. Evaluating effects of water withdrawals and impoundments on fish assemblages in southern New England streams, USA. Fisheries Management and Ecolog 17 (3), 272-283. doi: 10.1111/j.1365- 2400.2009.00724.x

Khedkar, Gulab Dattarao, Lutzky, Sigal, Rathod, Sandeep, Kalyankar, Amol, and David, Lior. 2014. A dual role of dams in fragmentation and support of fish diversity across the Godavari River basin in India. Ecohydrology 7 (6), 1560- 1573. doi: 10.1002/eco.1470.

133

Kundu, N., D. Mazumdar, and Homechaudhuri, S. 2014. Seasonal changes in fish assemblages and trophic guild in selected interconnected rivers in eastern India. Indian journal of Fisheries 61 (3), 35-44.

Lowe-McConnell, R. H. 1975. Fish communities in tropical freshwaters their distribution, ecology and evolution.: London.Pp. 337.

Lowe-McConnell, R. H. 1987. Ecological studies in tropical fish communities. Cambridge: Cambridge University Press, Cambridge.Pp 382.

Macnaughton, C. J. 2016. Ecological effects of flow and temperature regimes on fish communities in temperate rivers. PhD thesis submitted to the Department de sciences biologiques, Faculte des Arts et des Sciences, Universtite de Montreal. Pp. 193.

Matthews, M. J. 1998. Patterns in freshwater fish ecology. Massachusetts, USA.: Kluwer Academic publisher.Pp.670McKnight, Megan Wilde. 2007. Broad- scale patterns and determinants of beta diversity. A PhDthesis submitted to the University of North Carolina, Chapel Hill.

Mouillot, David, Graham, Nicholas A. J., Villéger, Sébastien, Mason, Norman W. H., and Bellwood, David R. 2013. A functional approach reveals community responses to disturbances Trends in Ecology & Evolution Vol. 28, : Elsevier. Pp. 167-177.

Penczak, T., Kruk, A., Głowacki, Ł, and Galicka, W. 2014. The role of streams of different sizes in recovery of fish fauna in the Pilica river catchment. River Research and Applications, 30 (8), 1034-1047. doi: 10.1002/rra.2699.

Pusey, Bradley J, Arthington, Angela H, and Read, Martin G. 1998. Freshwater fishes of the Burdekin River, Australia: biogeography, history and spatial variation in community structure. Environmental Biology of Fishes, 53 (3), 303-318. doi: 10.1023/A:1007468417947.

Raghavan, Rajeev, Das, Sandeep, Nameer, P. O., Bijukumar, A., and Dahanukar, Neelesh. 2016. Protected areas and imperilled endemic freshwater biodiversity in the Western Ghats Hotspot. Aquatic Conservation: Marine and Freshwater Ecosystems, 26, 78-90. doi: 10.1002/aqc.2653.

134

Ricklefs, R. E. and, and Dolph, Schluter. 1993. Species Diversity in Ecological Communities: University of Chicago Press.

Welcomme, R. L. 1985. River Fisheries. FAO Fisheries Technical Paper Pp. 330.

Winemiller, K. O. 2010. Preface: Stream fish communities from patch to Intercontinental convergences. In Community Ecology of Stream Fishes: Concepts, Approaches, and Techniques. American Fisheries Syposium 73. Keith B. Gido and Donald A. Jackson (ed). Pp 23-28.

Xu, J., Guohuan Su., Ying Xiong, M. Akasaka, J. G. Molinos, S. S.Matsuzaki, and M. Zhang 2015. Complimentary analysis of metacommunity nestednes and diversity partitioning highlights the need for a holistic conservation strategy for highland lake fish assemblages. Global Ecology and Conservation. V:288- 296. doi.org/10.1016/j.gecc0.2014.12.004

135

Appendix 2.1 Appendix - Ia. Details of sampling locations and disturbance regimes in the four sub- basins.

Ele vati River Stream Site Stream on Disturb. Disturb Disturbance basin name name order N E (m) rank level type Kadambi Bhadra 13.23 829 Bhadra joint 3 75.17 1 Low Bhadra Bhadra Bh7 3 13.23 75.17 860 1 Low Oldgrass Bhadra 854 Bhadra point 3 13.21 75.18 1 Low Bhadra Bhadra Kurinjikal 3 13.21 75.19 830 1 Low Khasigadd Bhadra 834 Bhadra e 3 13.21 75.19 4 Low Bhadra Bhadra Bhagvati 4 13.20 75.20 825 8 Medium checkdam Pandarmak Bhadra 824 Bhadra ki road 4 13.20 75.20 3 Low Bh- Bhadra kachige 778 Bhadra joint 4 13.21 75.23 1 Low Bhadra Bhadra Khagundi 5 13.21 75.24 761 2 Low Bhadra Bhadra KIOCL 5 13.21 75.25 770 1 Low Nagraj water extraction for Bhadra 760 Bhadra mane 5 13.21 75.25 1 Low Kudremukh town Kudremuk Bhadra 755 Bhadra h bridge 5 13.21 75.25 2 Low Hosmakki Bhadra 750 Bhadra joint 5 13.21 75.26 3 Low Bhadra Biligal Biligal1 4 13.21 75.22 775 3 Low Bhadra Biligal Biligal3 3 13.20 75.22 805 1 Low Bhadra Biligal Biligal5 3 13.20 75.22 816 1 Low Bhadra Kachige Kachige1 4 13.21 75.23 778 1 Low Bhadra Kachige Kachige3 4 13.20 75.23 785 1 Low Bhadra Kachige Kachige4 4 13.20 75.23 790 1 Low Bhadra Kachige Kachige5 4 13.19 75.23 806 1 Low Bhadra Singsar Singsar1 4 13.21 75.22 810 5 Medium Bhadra Singsar Singsar2 4 13.21 75.21 814 1 Low Bhadra Singsar Singsar3 4 13.22 75.21 928 1 Low Bhadra Kunya Kunya1 3 13.21 75.25 780 1 Low Bhadra Kunya Kunya2 3 13.21 75.24 782 1 Low Bhadra Kunya Kunya3 3 13.22 75.24 795 1 Low Bhadra Kunya Kunya4 3 13.22 75.24 824 1 Low Bhadra Kunya Kunya5 3 13.22 75.24 829 8 Medium checkdam water extraction for Carman Carman1 13.23 795 Bhadra 4 75.26 1 Low Kudremukh town Bhadra Carman Carman2 4 13.24 75.26 810 1 Low Bhadra Carman Carman3 3 13.24 75.25 814 1 Low Bhadra Carman Carman4 3 13.24 75.25 898 1 Low Hosmak Hosmakki 910 Bhadra ki 1 1 13.21 75.26 1 Low

136

Ele vati River Stream Site Stream on Disturb. Disturb Disturbance basin name name order N E (m) rank level type Hosmak Hosmakki 920 Bhadra ki 2 4 13.20 75.26 8 Medium checkdam Hosmak Hosmakki 930 Bhadra ki 3 4 13.20 75.26 1 Low Hosmak Hosmakki 767 Bhadra ki 4 3 13.21 75.28 1 Low Nelibed Nelibedu1 802 Bhadra u 3 13.20 75.28 1 Low Nelibed Nelibedu3 809 Bhadra u 3 13.20 75.28 8 Medium checkdam Nelibed Nelibedu5 861 Bhadra u 3 13.20 75.28 4 Low 105 Somvati Somvati1 Bhadra 2 13.15 75.30 3 1 Low private hydropower Somvati Somvati2 802 Bhadra 3 13.18 75.32 12 High plant Tunga Mudba Mudba1 4 13.31 75.13 664 1 Low Tunga Mudba Mudba2 4 13.31 75.13 670 1 Low Tunga Mudba Mudba4 4 13.31 75.12 675 1 Low Tunga Mudba Mudba5 3 13.31 75.12 700 1 Low Tunga Mudba Mudba8 2 13.30 75.13 700 5 Medium Mundsa Mundsar1 13.32 652 Tunga r 4 75.13 1 Low Mundsa Mundsar2 667 Tunga r 4 13.32 75.12 1 Low Mundsa Mundsar4 675 Tunga r 4 13.33 75.11 5 Medium 1waterpump Mundsa Mundsar8 700 Tunga r 4 13.32 75.12 1 Low Tunga Turad Turad1 4 13.30 75.15 642 1 Low Tunga Turad Turad2 4 13.30 75.16 689 1 Low Tunga Turad Turad3 4 13.31 75.16 690 5 Medium Tunga Turad Turad4 4 13.31 75.16 754 1 Low Karucha Karuchar1 667 Tunga r 4 13.33 75.14 5 Medium 1waterpump Karucha Karuchar2 672 Tunga r 4 13.33 75.13 5 Medium Karucha Karuchar3 686 Tunga r 4 13.33 75.13 2 Low Karucha Karuchar4 691 Tunga r 4 13.34 75.13 1 Low Karucha Karuchar5 693 Tunga r 4 13.34 75.13 5 Medium 1waterpump Karucha Karuchar6 697 Tunga r 4 13.34 75.18 5 Medium Tunga Korkan Korkan1 4 13.34 75.17 656 5 Medium Tunga Korkan Korkan2 4 13.34 75.17 661 5 Medium 1waterpump Tunga Korkan Korkan3 4 13.34 75.17 675 1 Low Tunga Korkan Korkan4 4 13.34 75.18 676 1 Low Tunga Korkan Korkan6 4 13.34 75.18 679 1 Low Tunga Korkan Korkan11 4 13.33 75.18 880 5 Medium

137

Ele vati River Stream Site Stream on Disturb. Disturb Disturbance basin name name order N E (m) rank level type Gangeh Ghole2 13.27 776 Tunga ole 3 75.16 1 Low Gangeh Ghole4 773 Tunga ole 4 13.28 75.16 1 Low Gangeh Ghole6 716 Tunga ole 4 13.29 75.16 1 Low Gangeh Ghole8 698 Tunga ole 3 13.29 75.16 5 Medium 1waterpump Tunga Muje Muje3 3 13.36 75.16 662 5 Medium 1waterpump Tunga Muje Muje4 3 13.36 75.16 667 5 Medium Tunga Muje Muje7 4 13.36 75.16 680 5 Medium Tunga Tanikod Tanikod2 2 13.36 75.20 736 5 Medium Tunga Tanikod Tanikod4 2 13.35 75.20 811 1 Low Mudba- Vimala mundsar 761 Tunga joint 4 13.32 75.13 2 Low Tunga Vimala Tn2 5 13.32 75.13 681 2 Low Acharmak Vimala 677 Tunga k 5 13.32 75.14 5 Medium Tumbahall Vimala 673 Tunga a 5 13.32 75.13 5 Medium Before Vimala Acharmak 670 Tunga ki 5 13.33 75.13 5 Medium Keshav Vimala 667 Tunga mane 5 13.33 75.14 5 Medium Before Vimala 665 Tunga Kerekatte1 5 13.33 75.14 5 Medium Before Vimala 665 Tunga Kerekatte2 5 13.33 75.14 5 Medium Tunga Tunga Kerekatte 5 13.33 75.14 656 10 High Kerekatte Tunga 655 Tunga school 5 13.33 75.14 6 Medium washing & bathing Trogen Tunga 650 Tunga point 5 13.33 75.15 5 Medium Before Tunga Anand 650 Tunga mane 5 13.34 75.16 2 Low Anand unga 650 Tunga mane 5 13.34 75.16 5 Medium 1waterpump Tunga Tunga Yadgar 5 13.34 75.17 650 1 Low Tunga Tunga Muje joint 6 13.35 75.17 645 5 Medium 1waterpump Tanikod Tunga 645 Tunga joint 6 13.36 75.20 1 Low Toursit Tunga 635 Tunga point 6 13.36 75.21 6 Medium 1 waterpump Tunga Tunga Salmara 6 13.37 75.21 630 5 Medium 1 waterpump Tunga Tunga Nemmar 6 13.38 75.21 630 11 High 1waterpump Malapr Malapra Amta 74.30 720 2 waterpump abha bha 3 15.71 5 Medium Malapr Malapra Torali 715 abha bha 3 15.71 74.32 5 Medium 1 waterpump

138

Ele vati River Stream Site Stream on Disturb. Disturb Disturbance basin name name order N E (m) rank level type 1 waterpump, Malapr Malapra Devachiha riverbed completely 720 abha bha tti dry during May 3 15.71 74.34 7 Medium 2014 Malapr Malapra Habbanhat 714 abha bha ti 3 15.72 74.36 6 Medium barrage Malapr Malapra Kusmali 700 abha bha 3 15.71 74.38 7 Medium Malapr Malapra Malavi 711 abha bha 3 15.66 74.43 5 Medium Malapr Malapra 1waterpump, Olmani 693 abha bha 3 15.71 74.41 12 High dynamiting Malapr Malapra Shankerpet 653 abha bha i 4 15.68 74.42 7 Medium washing & bathing Malapr Malapra sand mining, Asoga 643 abha bha 4 15.63 74.48 10 High waterpump,fishing Malapr Malapra Shedegali 650 abha bha 4 15.62 74.50 7 Medium washing & bathing sand mining, Malapr Malapra Rumewadi 655 waterpump,fishing, abha bha 5 15.63 74.52 9 High pollution Malapr Malapra fishing, waterpump, Kupatgiri 650 abha bha 5 15.64 74.53 9 High pollution Malapr Mangetr Katgali 74.47 713 abha i nala 4 15.72 5 Medium 1 waterpump Malapr Mangetr Valmiki 700 abha i nala 4 15.71 74.46 1 Low Malapr Haltar Shiroli 74.43 672 abha nala 4 15.55 7 Medium washing & bathing Mhadei Mhadei Degaon 3 15.54 74.35 685 1 Low Before Mhadei Mhadei 635 Kongla 5 15.59 74.38 1 Low Bhandu Mhadei Kongla 15.61 627 ra 4 74.39 5 Medium Kotni Mhadei 15.62 610 nadi Kotni 4 74.35 1 Low Below Mhadei Mhadei 48 Kishnapur 6 15.56 74.22 1 Low Mhadei Mhadei Ustem 6 15.56 74.21 44 5 Medium barrage Mhadei Mhadei Ustem 2 6 15.56 74.21 44 8 Medium sand-boulder Mhadei Mhadei Sonal 41 6 15.54 74.20 10 High mining sand-boulder Mhadei Mhadei Cudcem 31 6 15.53 74.17 10 High mining Mhadei Mhadei Velgeum 6 15.52 74.16 27 7 Medium barrage Mhadei Mhadei Khadaki 6 15.50 74.14 18 8 Medium barrage Mhadei Mhadei Khotode 6 15.49 74.13 14 7 Medium Mhadei Mhadei Waghurme 7 15.47 74.03 14 12 High bauxite mining Mhadei Mhadei Jamgaon 4 15.56 74.39 673 5 Medium Bhandu Mhadei Nerse 650 ra 4 15.60 74.41 1 Low Panshet Mhadei Talewadi 760 nala 3 15.55 74.32 1 Low

139

Ele vati River Stream Site Stream on Disturb. Disturb Disturbance basin name name order N E (m) rank level type Bail illegal fishing by Mhadei Bail nadi 721 nadi 3 15.64 74.28 9 High locals Mhadei Kalasa Satrem 3 15.62 74.22 112 5 Medium Polluted water Mhadei Kalasa Derodem 4 15.60 74.22 84 5 Medium waterpump house Mhadei Kalasa Nanodem 62 by Goa Irrigation 4 15.58 74.21 5 Medium dept Mhadei Kalasa Kankumbi 3 15.69 74.21 750 9 High illegal fishing water extraction for Mhadei Kalasa Delta hotel 694 3 15.68 74.19 7 Medium hotel Mhadei Kalasa Checkpost 3 15.68 74.18 733 1 Low Kotryac Mhadei Thane 15.61 85 hi nadi 5 74.15 5 Medium fishing Kotyrac Mhadei Hedode 79 hi nadi 5 15.57 74.14 5 Medium Kotryac Mhadei Naneli 65 hi nadi 5 15.58 74.14 8 Medium barrage Kotryac Mhadei Velus 49 hi nadi 5 15.55 74.14 9 High barrage Mhadei Patwal Patwal 4 15.49 74.15 12 5 Medium Vasant Mhadei Ragada 135 bandhara 1 15.40 74.26 8 Medium checkdam Mhadei Ragada Jambolim 4 15.39 74.22 21 5 Medium Mhadei Ragada Satpali 4 15.40 74.21 61 7 Medium vehicle washing waterpump for Mhadei Ragada Panas 32 4 15.41 74.18 7 Medium oilpalm Mhadei Ragada Shivade 4 15.42 74.14 16 5 Medium Mhadei Ragada Murmune 5 15.47 74.12 13 5 Medium Dudhsa Dudhsagar Mhadei 146 gar fall 4 15.31 74.31 3 Low washing & bathing Dudhsa Devachi Mhadei 97 gar kon 4 15.33 74.28 1 Low tourist vehicles Dudhsa Dudhsagar Mhadei 85 frequently passt gar juntion. 4 15.34 74.26 3 Low hrough vehicle washing, Dudhsa Mhadei Cullem 74 bathing, water gar 5 15.34 74.25 7 Medium extraction for hotels Dudhsa Mhadei Shigaon 58 gar 5 15.34 74.21 5 Medium Dudhsa Mhadei Dabal 22 gar 5 15.36 74.13 5 Medium Caranzh Mhadei Cumtol 1 104 ol 3 15.50 74.22 1 Low Caranzh washing clothes, Mhadei Cumtol 2 50 ol 4 15.52 74.20 7 Medium cashew liquer plant Karanjh Mhadei Karanjhol 83 ol 4 15.35 74.28 1 Low

140

Appendix 1b. Systematic list of freshwater fishes found in the study area.

Abunda Orders Family Species English names nce IUCN Osteolossiform es Notopteridae Notopterus notopterus Bronze featherback 2 LC Cypriniformes Cyprinidae Barilius bakeri Malabar baril 995 LC Barilius barna Barna baril 1 LC EN B1 abiii Barilius canarensis Jerdon’s baril 291 +2abiii Barilius bendelisis Hamilton’s baril 30 LC Bariliius spp - 4 NE Barbodes wynaadensis Wayanaad mahseer 184 CR A2ace Cirrihinus fulungee Deccan white carp 432 LC Cirrhinus spp - 3 NE Chela cachius Silver hatchet chela 1 LC Dawkinsia arulius Aruli barb 171 EN B2abiii Dawkinsia filamentosa Filament barb 453 LC Devario spp - 5 NE Devario malabaricus Giant danio 3494 LC Garra bicornuta Tunga garra 185 NT Garra mullya Mullya garra 3013 LC Garra stenorhynchus Nilgiri garra 24 LC Haludaria melanampyx Melon barb 55 DD Hypselobarbus curmuca Curmuca barb 204 EN A2acd Hypselobarbus dobsoni Krishna carp 56 DD Hypselobarbus dubius Nilgiri barb 1 EN B2abiii Hypselobarbus jerdoni Jerdon’s carp 234 LC Hypselobarbus CR B2ab thomassi Red canarese barb 1 iii Fringed-lipped peninsula Labeo fimbriatus carp 17 LC Labeo porcellus Bombay labeo 2 LC Labeo spp - 6 NE Oreichthys cosuatis Kosuati barb 2 LC Osteochilichthys nashii Nash's barb 828 LC Osteochilichthys thomassi Konti barb 13 LC Pethia narayani Narayan barb 381 LC Pethia sp - 1 NE Pethia punctata Dotted sawfin barb 2 LC VU B2 Pethia setnai Indego barb 205 abiii Pethia striata Kudremukh barb 24 NE Pethia ticto Ticto barb 16 LC Puntius amphibius Scarlet-banded barb 480 DD Puntius chola Chola barb 4 LC Puntius dorsalis Long snouted barb 5 LC Puntius sahyadriensis Khavli barb 441 LC Puntius sophore Spotfin swamp barb 21 LC Rohtee ogilbii Vatani rohtee 8 LC

141

Abunda Orders Family Species English names nce IUCN Rasbora daniconius Slender rasbora 1161 LC Rasbora labiosa Slender rasbora 12 LC Salmophasia bacaila Large razor belly minnow 67 LC Salmophasia boopis Boopis razer belly minnow 3182 LC Novacula razor belly Salmophasia novacula minnow 93 LC Schismatorhynchos EN A2acd nukta Nukta 1 +3acd Systomus sarana sarana Olive barb 64 LC Systomus sarana subnastus - 27 NE EN Tor khudree Deccan mahseer 119 B2acde Paracanthocobitius Balitoridae mooreh Cobitis mooreh 33 NE Bhavania australis Western ghat loach 19 LC VU B2ab Balitora mysorensis Slender stone loach 1 iii Nemacheilus denisoni - 24 LC Nemacheilus spp - 24 NE Mesonemacheilus spp - 1 NE Nemacheilus triangularis Zodiac loach 4 LC Nemacheilus rueppelli Mongoose loach 5 LC Nemacheilus semiarmatus Dotted loach 3 LC Lepidocephalus thermalis Common spiny loach 31 LC EN B2 Cobitidae Botia striata Zebra loach 10 abiii EN B1abiii Siluriformes Bagridae Batasio sharavatiensis Sharavati batasio 8 +2abiii Mystus armatus Kerala mystus 11 LC Mystus bleekeri Day’s mystus 101 LC Mystus cavacius Gangetic mystus 90 LC Mystus gulio Long whiskered catfish 202 LC Mystus keletius Yellow catfish 67 LC Siluridae Ompok bimaculatus Butter catfish 8 NT Ompok malabaricus Goan catfish 15 LC Schilibidae Eutropiichthys spp - 1 NE Sisoridae Glyptothorax spp2 - 1 NE Glyptothorax spp3 - 1 NE Glyptothorax spp4 - 1 NE Claridae Clarias batrachus Walking catfish 1 LC Mugiliformes Mugilidae Migul spp - 9 NE Beloniformes Belonidae Xenentodon cancila Freshwater garfish 93 LC Cyprinodontifo Aplocheilida rmes e Aplocheilus lineatus The striped panchax 59 LC Synbranchifor Mastacemeli Mastacemeblus mes dae armatus Spiny eel 1 LC Perciformes Ambassidae Chanda nama Elongated glass perchlet 112 LC Parambassis ranga Indian glassy fish 63 LC Western ghat glassy Parambassis thomassi perchlet 17 LC Cichlidae Etroplus maculatus Orange chromid 21 LC

142

Abunda Orders Family Species English names nce IUCN Etroplus suratensis Green chromid 155 LC Gobidae Glossogobius giuris Bareye goby 33 LC Glossogobius spp - 6 NE Sicyopterus griseus Clown goby 24 LC Channidae Channa gachua Dwarf snakehead 4 LC Channa micropeltes Giant snakehead 1 LC Channa marulius Murrel 2 LC Channa punctata Snakehead 1 LC Channa striata Snakehead murrel 25 LC Tetraodontifor Tetraodontid Carinotetradon Dwarf pufferfish 11 VU mes ae travancoricus Arothron leopardus Banded leopardblowfish 2 DD

Note: LC – Least concern, NE – Not evaluated, DD-data deficient, EN-Endangered, CR- Critically endangered, VU – Vulnerable, NT- Near thrtn.

Appendix – 2.2: .Range (minimum, maximum), mean and standard deviation of environmental variables.

Variable Unit Min Max Mean SD Total alkalinity Mg/l 11 125.5 48.60 23.98 Free.co2 Mg/l 0.46 81.73 41.75 25.09 Chloride Mg/l 8.52 284 26.08 29.20 Calcium hardness Mg/l 3.69 426.72 50.52 86.69 Total hardness Mg/l 7.3 904 68.76 124.6 Inorganic nitrates Mg/l 0.01 1.45 0.25 0.22 Inorganic phosphates Mg/l 0.001 0.77 0.14 0.12 Dissolved oxygen Mg/l 4.55 16.62 8.90 2.10 Velocity m3/s 0.029 1.3 0.23 0.22 Discharge m3/s 0 0.019 0.006 0.006 Depth- width ratio - 0.005 0.16 0.040 0.027 0 Water temperature C 16 34.2 23.21 3.90 pH - 5.6 18.6 7.42 1.31 Electrical conductivity µm/s 0 0.16 0.03 0.03 Total dissolved solids ppt 0 0.12 0.027 0.026 Canopy cover % 0 95 49.10 25.50 Rocks % 0 60 11.20 15.92 Boulders % 0 80 11.19 12.11 Cobbles % 0 50 16.08 11.46 Pebbles % 0 50 18.43 12.38 Gravels % 0 80 28.42 17.05 Sand % 0 60 11.50 10.67 Leaves & woods % 0 90 3.99 9.33

143

Appendix – 2.3. Exploratory plots

Figure i). Rank abundance of fish species in four river sub-basins (a) Bhadra, (b) Tunga, (c) Malaprabha, (d) Mhadei.

144

Figure ii). Dominant fish families in each river sub-basin respectively. (a) Bhadra (b) Tunga (c) Malaprabha (d) Mhadei.

145

Figure iii). Size class for four dominant fish species in regulated and non-regulated basins. (a) Rasbora daniconius (b) Salmophasia boopis (c) Devario malabaricus (d) Garra mullya.

146

Figure iv). Size class for four dominant fish species in studied river sub-basins.(a) Rasbora daniconius (b) Salmophasia boopis (c) Devario malabaricus (d) Garra mullya.

147

Figure v). (a) Relationship between species richness and habitat covariates.

148

Figure vi). (b) Relationship between species richness and habitat covariates.

149

Figure vii). (c) Relationship between species abundance and habitat covariates.

150

Figure viii). (d) Relationship between species abundance and habitat covariates.

151

Appendix – 2.4

Figure ix). The diagnostic plots for generalized linear model for fish richness and fish abundance (a) residual fish richness vs fitted fish richness (b) residual fish abundance vs fitted fish abundance

152

Appendix 3.1 Freshwater fish species encountered in both Bhadra and Tunga sub-basins in Kudremukh National Park.

Total Common Unique to Unique to Bhadra Tunga species species Bhadra Tunga species species Paracanthocobitius Garra Paracanthocobitius Paracanthocobitius mooreh Barilius bakeri stenorhynchus mooreh Barilius bakeri mooreh Aplocheilus Barilius Aplocheilus Barilius Aplocheilus lineatus bendelisis lineatus bendelisis lineatus Barilius Barilius Balitora mysorensis canarensis Balitora mysorensis canarensis Balitora mysorensis Barbodes Barbodes Barilius bakeri wynaadensis Chanda nama wynaadensis Barilius bakeri Batasio Batasio Barilius bendelisis sharavatiensis Cirrhinus spp sharavatiensis Barilius canarensis Bhavania Bhavania Barilius canarensis australis Glyptothorax sp3 australis Barilius bendelisis Barbodes Hypselobarbus Barilius wynaadensis Botia striata dobsoni Botia striata wynaadensis Batasio Hypselobarbus Batasio sharavatiensis Channa striata jerdoni Channa striata sharavatiensis Dawkinsia Mastacemeblus Dawkinsia Bhavania australis arulius armatus arulius Bhavania australis Devario Devario Botia striata malabaricus Mystus gulio malabaricus Botia striata Garra Nemacheilus Garra Chanda nama bicornuta denisoni bicornuta Chanda nama Nemacheilus Channa striata Garra mullya semiarmatus Garra mullya Channa striata Hypselobarbus Nemacheilus Garra Cirrhinus spp curmuca triangularis stenorhynchus Cirrhinus spp Osteochilus Ompok Hypselobarbus Dawkinsia arulius nashii malabaricus curmuca Dawkinsia arulius Devario Puntius Osteochilus Devario malabaricus sahyadriensis Parambassis ranga nashii malabaricus Rasbora Parambassis Puntius Garra bicornuta daniconius thomassi sahyadriensis Garra bicornuta Salmophasia Rasbora Garra mullya boopis Pethia setnai daniconius Garra mullya Garra Salmophasia stenorhynchus Tor khudree Pethia striata boopis Glyptothorax sp3 Xenentodon Hypselobarbus Glyptothorax sp3 cancila Pethia ticto Tor khudree curmuca Hypselobarbus Salmophasia Xenentodon Hypselobarbus curmuca novacula cancila dobsoni Hypselobarbus Hypselobarbus dobsoni jerdoni Hypselobarbus Mastacemeblus jerdoni armatus Mastacemeblus armatus Mystus gulio Nemacheilus Mystus gulio denisoni Nemacheilus Nemacheilus denisoni triangularis N. semiarmatus N. semiarmatus

153

Total Common Unique to Unique to Bhadra Tunga species species Bhadra Tunga species species Nemacheilus Ompok triangularis malabaricus Ompok malabaricus Osteochilus nashii Osteochilus nashii Parambassis ranga Parambassis Parambassis ranga thomassi Parambassis thomassi Pethia setnai Pethia setnai Pethia striata Pethia striata Pethia ticto Puntius Pethia ticto sahyadriensis Puntius sahyadriensis Rasbora daniconius Salmophasia Rasbora daniconius boopis Salmophasia Salmophasia boopis novacula Salmophasia novacula Tor khudree Tor khudree Xenentodon cancila Xenentodon cancila

154

Appendix 3.2 (a) Species dissimilarity with respect to combined basin reference maximum across Bhadra and Tunga and (b) species dissimilarity with respect to basin reference maximum for endemic species using incidence data. Species dissimilarity is higher in Bhadra than Tunga in both (a) & (b) panel.

155

Appendix 3.3 (a) Species dissimilarity with respect to the individual basin reference maximum using incidence data for all species and (b) for endemic species. Species dissimilarity is consistently higher in Tunga than Bhadra basin.

156

Appendix 4.1 Supporting Information on catchment details across four river sub-basins S1 A Table. Catchment area ______River Area Direction Basin State Sub-basin (km2) of flow Mhadei 425.96 West Mhadei Karnataka & Goa Malaprabha 744.32 East Krishna Karnataka Bhadra 225.12 East Krishna Karnataka Tunga 160.00 East Krishna Karnataka

S1 B Table. Details of environmental variables measured in the study. ______Variables Units Measurement details Site characteristics Elevation Location (Lat. Long) Degree decimal GPS etrex Garmin Water quality variables Water temperature 0C Thermometer Electrical conductivity µms-1 Hanna Instruments - 98302 Total dissolved solids parts per trillion Hanna Instruments - 98304 pH --- Hanna Instruments - 96107 Dissolved oxygen mgL-1 Winkler’s titration method Inorganic nitrates mgL-1 Spectro-photometry Inorganic phosphates mgL-1 Spectro-photometry -1 Free CO2 mgL water analysis in the lab Calcium hardness mgL-1 water analysis in the lab Total hardness mgL-1 water analysis in the lab Total alkalinity mgL-1 water analysis in the lab

Stream characteristics Stream order 1-5 Survey of India toposheet (1:50,000) Stream width m meter by using a measuring tape Depth m average 3 depth readings along a stream width Canopy cover % Visual estimation Rocks % Boulders % Cobbles % Pebbles % Gravels % Sand % Leaves and woods %

157

Supporting Information - S2 Table1. Systematic list of fish species with multiple guild definitions of freshwater fish species sampled across regulated (RSB) and non- regulated sub-basins (NRSB) in the study area. ______Order, Family, Species Sub-basins Basis of guild definition Water column Feeding Flow Life history Reproductive Position Preferences response cues strategy ______Cyprinodontiformes Aplocheilidae Aplocheilus lineatus TN, ML SD Hetertrophic Rheophilic Opportunistic Lithophils Beloniformes Belonidae Xenentodon cancila BH, TN SD Hetertrophic Rheophilic Intermediate Phytophils Cypriniformes Balitoridae Paracanthocobitius mooreh * TN, ML BD Hetertrophic Rheophilic Periodic Lithophils Balitora mysorensis* BH, TN BD Phytophagous Rheophilic Periodic Lithophils Bhavania australis* BH, TN BD Omnivore Rheophilic Periodic Psammophils Mesonemacheilus sp* TN BD Phytophagous Rheophilic Periodic Lithophils Nemacheilus denisoni TN, ML, MH BD Omnivore Rheophilic Periodic Lithophils Nemacheilus rueppelli* ML BD Omnivore Rheophilic Periodic Lithophils Nemacheilus triangularis* BH, TN BD Phytophagous Rheophilic Periodic Phytophils Nemacheilus semiarmatus* TN BD Phytophagous Rheophilic Periodic Lithophils Lepidocephalus thermalis* ML, MH BD Phytophagous Rheophilic Periodic Lithophils Cobitidae Botia striata BH, TN BD Omnivore Rheophilic Periodic Psammophils Cyprinidae Barilius bakeri* BH, TN SD Omnivore Rheophilic Periodic Lithopelagophils Barilius barna TN SD Omnivore Rheophilic Periodic Lithopelagophils

158

Order, Family, Species Sub-basins Basis of guild definition

Water column Feeding Flow Life history Reproductive

Position Preferences response cues strategy

Barilius canarensis* BH, TN SD Omnivore Rheophilic Periodic Lithopelagophils Barilius bendelisis BH, TN SD Omnivore Rheophilic Periodic Lithopelagophils Barilius sp TN SD Omnivore Rheophilic Periodic Lithopelagophils Barbodes wyanaadensis* BH, TN MCD Omnivore Rheophilic Periodic Psammophils Cirrhinus fulungee ML MCD Omnivore Rheophilic Periodic Lithophils Cirrhinus sp* TN MCD Omnivore Rheophilic Periodic Lithophils Chela cachius ML MCD Omnivore Rheophilic Periodic Phytolithophils Devario malabaricus* BH, TN, ML, MH MCD Heterotrophic Rheophilic Opportunistic Phytolithophils Dawkinsia arulius* BH, TN MCD Omnivore Rheophilic Intermediate Phytophils Dawkinsia filamentosa* MH MCD Omnivore Rheophilic Intermediate Phytophils Garra bicornuta* BH, TN, ML, MH BD Phytophagous Rheophilic Opportunistic Lithopelagophils Garra mullya* BH, TN, ML, MH BD Phytophagous Rheophilic Opportunistic Lithopelagophils Garra stenorhynchus* BH, ML, MH BD Phytophagous Rheophilic Opportunistic Lithopelagophils Hypselobarbus curmuca* BH, TN, ML BD Omnivore Rheophilic Periodic Phytophils Hypselobarbus dobsoni* TN, ML BD Phytophagous Rheophilic Periodic Lithophils Hypselobarbus dubius* ML BD Omnivore Rheophilic Periodic Lithophils Hypselobarbus jerdonii* BH, TN MCD Phytophagous Rheophilic Periodic Lithophils Hypselobarbus thomassi* BH, TN MCD Phytophagous Rheophilic Periodic Lithophils Labeo fimbriatus* ML BD Phytophagous Limnophilic Periodic Lithophils Labeo porcellus* ML BD Phytophagous Limnophilic Periodic Lithophils Labeo spp* TN BD Phytophagous Limnophilic Periodic Lithophils Oreichthys cosuatis ML MCD Heterotrophic Rheophilic Periodic Phytophils Osteochilius nashii* BH, TN, ML MCD Omnivore Rheophilic Periodic Lithophils Pethia sp* ML MCD Heterotrophic Rheophilic Intermediate Phytophils

159

Order, Family, Species Sub-basins Basis of guild definition

Water column Feeding Flow Life history Reproductive

Position Preferences response cues strategy Pethia setnai* TN, ML, MH MCD Heterotrophic Rheophilic Intermediate Phytophils Pethia striata* TN BD Phytophagous Rheophilic Periodic Phytophils

Pethia ticto TN, ML MCD Omnivore Eurytopic Periodic Phytophils Puntius amphibius* ML, MH MCD Heterotrophic Rheophilic Intermediate Phytophils Puntius chola ML MCD Omnivore Rheophilic Intermediate Phytophils Puntius dorsalis ML MCD Phytophagous Rheophilic Intermediate Phytophils Puntius sahyadriensis* BH, TN, ML MCD Omnivore Rheophilic Intermediate Phytophils Puntius sophore ML MCD Heterotrophic Eurytopic Periodic Phytophils Rasbora daniconius BH, TN, ML, MH MCD Heterotrophic Eurytopic Opportunistic Phytophils Rasbora labiosa ML, MH MCD Omnivore Eurytopic Opportunistic Phytophils Rhohtee ogilbii* TN MCD Phytophagous Rheophilic Periodic Phytophils Salmophasia bacaila ML, MH SD Omnivore Limnophilic Periodic Phytophils Salmophasia boopis BH, TN, ML, MH SD Heterotrophic Limnophilic Periodic Phytophils Salmophasia novacula* ML SD Heterotrophic Limnophilic Periodic Phytophils Schismatorhynchus nukta ML BD Phytophagous Rheophilic Periodic Lithophils Systomus sarana sarana ML MCD Phytophagous Limnophilic Intermediate Phytophils Systomus sarana subnastus ML MCD Phytophagous Limnophilic Intermediate Phytophils Tor khudree* BH, TN, ML, MH MCD Omnivore Rheophilic Periodic Lithophils Mastacembeliformes Mastacembelidae Mastacembelus armatus BH, TN BD Heterotrophic Rheophilic Equilibrium Phytophils Osteoglossiformes Notopreridae Notopterus notopterus TN BD Omnivore Rheophilic Intermediate Lithophils

160

Order, Family, Species Sub-basins Basis of guild definition

Water column Feeding Flow Life history Reproductive

Position Preferences response cues strategy Perciformes Ambassidae Parambassis ranga TN, ML MCD Heterotrophic Eurytopic Intermediate Phytophils Parambassis thomassi* TN MCD Heterotrophic Eurytopic Intermediate Phytophils Chanda nama TN MCD Heterotrophic Eurytopic Intermediate Pelagophils Channidae Channa gachua ML BD Heterotrophic Eurytopic Equilibrium Pelagophils Channa micropeltis ML BD Heterotrophic Eurytopic Equilibrium Pelagophils Channa punctata ML BD Heterotrophic Eurytopic Equilibrium Pelagophils Channa striata BH, TN, ML, MH BD Heterotrophic Eurytopic Equilibrium Pelagophils Cichlidae Etroplus suratensis ML BD Omnivore Eurytopic Equilibrium Lithophils Gobidae Glossogobius guirus guirus ML BD Heterotrophic Eurytopic Opportunistic Lithophils Sisoridae Glyptothorax sp2* BH BD Heterotrophic Rheophilic Periodic Lithophils Glyptothorax sp3* TN BD Heterotrophic Rheophilic Periodic Lithophils Glyptothorax sp4* ML BD Heterotrophic Rheophilic Periodic Lithophils Siluriformes Bargidae Batasio sharavatiensis* BH, TN BD Heterotrophic Eurytopic Periodic Phytophils Mystus bleekeri ML BD Heterotrophic Eurytopic Equilibrium Phytophils Mystus cavacius ML BD Heterotrophic Rheophilic Periodic Phytophils Mystus gulio TN, ML BD Heterotrophic Rheophilic Periodic Phytophils Mystus keletius* ML BD Heterotrophic Rheophilic Periodic Phytophils

161

Order, Family, Species Sub-basins Basis of guild definition

Water column Feeding Flow Life history Reproductive

Position Preferences response cues strategy Claridae Clarias batrachus ML BD Omnivore Eurytopic Opportunistic Phytophils Siluridae Ompok bimaculatus ML BD Heterotrophic Eurytopic Periodic Polyphils

Ompok malabaricus* TN BD Heterotrophic Eurytopic Periodic Polyphils

Schilibidae Eutropiichthys sp* ML BD Heterotrophic Rheophilic Periodic Phytophils Note: SD – Surface dwellers, MCD – Mid-column dwellers, BD – Bottom dwellers, * indicate species endemic to the Western Ghats region Regulated sub-basins: Malaprabha (ML), Mhadei (MH); Non-regulated sub-basins: Tunga (TN), Bhadra (BH).

162

Table 2. Literature review for fish guilds sampled in the study period.

Guilds Guild definitions References cited

Water column position based Bottom dwellers – commonly [59]. Lowe-McConnell RH. guild found at the bottom surface, (1975). Fish communities in Mid-column dweller – found in tropical freshwaters their the mid-surface, Surface dweller distribution, ecology and – found near the surface evolution. London.

[60]. Lowe-McConnell RH. (1987). Ecological studies in tropical fish communities. Cambridge University Press, Cambridge, Cambridge. Feeding preference guilds Heterotrophic consist of [25]. Bhat A. (2002). A study of insectivores & carnivores, feed the diversity and ecology of on benthic insect, snails and Freshwater ﬁshes of four river small fish. Omnivore feed on systems of the Uttara Kannada variety of food items such as District, Karnataka, India. 178. food, leaves, insects etc. Phytophagous fishes consist of algivore, herbivore, detritivore [24]. Daniels RJR. (2002). feed largely on algae, Freshwater fishes of Peninsular fallen/decayed fruits & leaves. India. University Press India (Pvt) Ltd, Hyderabad.

[29]. Johnson J,Arunachalam M. (2012). Feeding habit and food partitioning in a stream fish community of Western Ghats, India. Environmental Biology of Fishes 93: 51-60.

[27]. Weliange WS,Amarsinghe US. (2007). Relationship between body shape and food habits of fish from three reservoirs of Sri Lanka. Asian Fisheries Science 20: 257-270.

[30]. Chakrabarty M, and S. Homechaudhuri. (2013). Fish guild structure along a longitudinally-determined ecological zonation of Teesta, an eastern Himalayan river in West Bengal, India. Arxius de Miscel-lania Zoologica 11: 196- 213

Flow-response guilds Eurytopic means all life stages [62]. Aarts BGW, Nienhuis PH. can occur both in lotic and lentic (2003). Fish zonations and waters. Limnophilic consists of guilds as the basis for all life stages are confined to assessment of ecological lentic water with macrophytes integrity of large rivers. and rheophilic means that all life Hydrobiologia 500:157-178 stages are confined to main river 163

channel. [30]. Chakrabarty M, and S. Homechaudhuri. (2013). Fish guild structure along a longitudinally-determined ecological zonation of Teesta, an eastern Himalayan river in West Bengal, India. Arxius de Miscel-lania Zoologica 11: 196- 213

[45]. Das MK, Sharma AP, Vass KK, Tyagi RK, Suresh VR, Naskar M,Akolkar AB. (2013). Fish diversity, community structure and ecological integrity of the tropical River Ganges, India. Aquatic Ecosystem Health & Management 16: 395-407.

Life-history cue-related guilds Equilibrium strategists are fishes [63].Welcomme RL. 1985. with small to medium body size, River Fisheries. FAO. 330. small clutch, high parental care & juvenile survivorship with intermediate maturity, [64]. Welcomme RL, opportunistic includes small Winemiller KO,Cowx IG. body size, early maturity, (2006). Fish environmental continuous reproduction, small guilds as a tool for assessment clutch size & little parental care, of ecological condition of rivers. periodic /seasonal strategists River Research and include fishes with large body Applications 22: 377-396. size, high fecundity, low juvenile survivorship, no parental care and late maturation [30]. Chakrabarty M, and S. and intermediate strategists Homechaudhuri. (2013). Fish include fishes with small body guild structure along a size, prolonged and distinct longitudinally-determined seasonal reproduction ecological zonation of Teesta, an eastern Himalayan river in West Bengal, India. Arxius de Miscel-lania Zoologica 11: 196- 213

Reproductive strategy guilds Lithophils includes rock & [62]. Aarts BGW, Nienhuis PH. gravel spawners, (2003). Fish zonations and lithopelagophils means rock, guilds as the basis for gravel spawners with pelagic assessment of ecological larvae, Phytolithophils means integrity of large rivers. non-obligatory plant spawners, Hydrobiologia 500:157-178 phytophils means obligatory plant spawners, Pelagophils [30]. Chakrabarty M, and S. means pelagic spawners, Homechaudhuri. (2013). Fish polyphils includes obligatory guild structure along a plant and psammophils consist longitudinally-determined of sand spawners. ecological zonation of Teesta, an eastern Himalayan river in West Bengal, India. Arxius de Miscel-lania Zoologica 11: 196-

164

213

Supporting Information - S3. S3 Figure 1. Principal Components Analysis for selection of environmental variables for further regression analyses: (A) Scree plot showing proportion of variance explained for the first four Principal Components and (B) Bi-plot showing correlations between different variables along PC axes.

165

PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 PC11 PC12 PC13 PC14 PC15 PC16 PC17 Tot.alk -0.04 -0.15 0.04 -0.07 0.09 -0.26 0.12 0.29 0.61 -0.58 0.13 -0.08 -0.12 0.24 -0.02 0.00 0.00 Free.co2 -0.21 -0.67 0.47 0.22 0.23 0.06 0.25 0.23 -0.17 0.11 -0.18 -0.01 0.01 -0.05 0.01 0.00 0.00 Chloride -0.02 -0.02 0.04 0.02 -0.07 -0.02 -0.05 0.14 0.07 -0.24 0.28 0.33 0.35 -0.76 0.14 0.01 0.00 Cal.hard 0.06 0.09 -0.06 -0.08 0.04 -0.11 -0.04 0.01 -0.24 -0.44 -0.60 -0.49 0.26 -0.21 0.04 0.00 0.00 Tot.hard 0.00 -0.11 0.04 0.14 -0.10 0.38 0.44 -0.70 0.12 -0.30 0.11 -0.06 0.10 0.01 0.00 0.00 0.00 Inorg.N 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 -0.03 0.00 0.01 -0.02 -0.17 -0.98 -0.07 -0.03 Diss.oxy 0.01 -0.02 0.02 0.02 -0.01 -0.08 -0.03 0.06 -0.01 0.06 0.11 0.09 0.87 0.45 -0.10 -0.02 0.00 Discharge 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 0.14 -0.94 DW.ratio 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 -0.07 0.99 0.13 Wat.temp 0.00 -0.20 -0.11 -0.28 0.01 -0.09 -0.06 -0.08 -0.55 -0.45 0.01 0.52 -0.16 0.23 -0.03 0.00 0.00 TDS 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -0.01 0.02 0.32 Cancov 0.93 -0.06 0.34 0.00 0.03 -0.06 0.00 -0.02 -0.01 0.00 0.03 0.03 -0.02 0.00 0.00 0.00 0.00 Rocks 0.00 -0.02 0.04 0.06 0.07 0.29 -0.22 0.19 -0.35 -0.20 0.62 -0.51 -0.04 0.05 -0.01 0.00 0.00 Boulders 0.11 -0.37 -0.27 -0.55 0.27 0.50 -0.23 0.00 0.25 0.11 -0.10 -0.02 0.09 -0.06 0.01 0.00 0.00 Cobbles 0.22 -0.26 -0.61 0.06 -0.32 -0.05 0.51 0.31 -0.12 0.12 0.06 -0.10 0.00 -0.04 0.00 0.00 0.00 Pebbles -0.03 -0.41 -0.12 -0.11 0.05 -0.63 -0.21 -0.45 0.01 0.16 0.23 -0.25 0.02 -0.12 0.01 0.00 0.00 Gravels -0.13 0.01 0.42 -0.62 -0.60 -0.02 0.17 0.06 -0.02 0.09 0.05 -0.14 0.01 -0.01 0.00 0.00 0.00 Sand -0.02 0.32 0.03 -0.36 0.62 -0.16 0.54 0.01 -0.13 0.11 0.17 -0.04 0.04 -0.08 0.01 0.00 0.00

166

Appendix – 5.1 Copeia

Published by: The American Society of Ichthyologists and Herpetologists « previous article : next article » Select Language ▼ translator disclaimer

Copeia 103(2):290-296. 2015 https://doi.org/10.1643/OT-12-172

A New Species of Pethia from the Western Ghats, India (Teleostei: Cyprinidae)

V. M. Atkore , J. D. Marcus Knight , K. Rema Devi , and J. Krishnaswamy

Received: December 18, 2012; Accepted: December 12, 2014

;

Published: May 4, 2015

[+] Author & Article Info

Pethia striata, new species, is described from the Tunga River in Kudremukh National Park, in the central part of the Western Ghats, Karnataka State, India. The new species is distinguished from its congeners by the combination of the following characters: absence of barbels; stiff and serrated last unbranched dorsal-fin ray; complete lateral line with 20–21 pored scales and a relatively small humeral spot one scale below the fourth lateral-line scale; a large black blotch covering lateral-line scales 17–19. In addition, the outer edges of body scales are dark, producing a striped pattern along the sides of the body. Pethia striata, new species, is presently known only from headwater-streams of the Tunga River basin.

167

Appendix – 5.2 River Research and Applications

RESEARCH ARTICLE

Assessing the recovery of fish assemblages downstream of hydrological barriers in India's Western Ghats

Authors V. Atkore., N. Kelkar and J. Krishnaswamy

 First published: 19 June 2017Full publication history  DOI: 10.1002/rra.3163 View/save citation  Cited by (CrossRef): 0 articles Check for updates

Abstract

River flow regulation by dams and barrages threatens freshwater fish diversity globally. However, factors contributing to the recovery of fish communities downstream of barriers to river flow are not well understood. It is crucial to identify processes that might enable river restoration despite the presence of river barriers. In this study, we assess recovery of fish species, including endemics, downstream of large and small barriers in the Malaprabha basin in the Western Ghats of India. We define “fish species recovery” as the proportion of fish species occurring in river reaches downstream of barriers, of the species pool occurring in upstream unregulated segments with similar elevation, stream order, and habitat characteristics. As per the serial discontinuity concept, we predicted that recovery will reduce immediately after, but gradually increase with, increasing distance downstream of barriers, due to contributions from unregulated streams joining the river. As expected, fish recovery decreased immediately downstream of barriers and increased at greater distances and declined when the number of upstream barriers increased, indicating cumulative impacts. Dissolved oxygen and total alkalinity were positively and negatively correlated with both recovery and distance from barrier. Water temperature and rocky instream habitat influenced recovery positively, but independent of distance from barriers. Recovery of fish species, including Western Ghats endemics, was promising even under the current level of river regulation in the area, mainly due to connectivity with undammed tributaries. Strict limits on future stream regulation within already regulated basins will be critical for conservation of freshwater fish biodiversity in this region.

168